dataset_name stringlengths 8 32 | series_description stringlengths 1.32k 2.25k | algorithm stringclasses 8 values | hyperparameters stringclasses 93 values |
|---|---|---|---|
3001-7.csv | A multivariate classification time-series dataset consists of 576 samples and 1 features with 1 numerical and 0 categorical features. Each instance has a window length of 3. The dataset has a sampling rate of 480.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00
The target column has 4 classes with entropy value 1.27 showing a Unbalanced dataset. Among the 576 samples the target ground-truth class has changed 377 times representing a percentage of 66.84%. There are 1 features in the dataset
Among the numerical predictors, the series has 0 numerical features detected as Stationary out of the 1 numerical features using the dickey-fuller test and the rest are Unstationary. 0 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.032. to 0.032 and kurtosis values of 0.83 to 0.83. The fractal dimension analysis yields values ranging from -1.40 to -1.40 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 1.00, maximum 1.00, mean 1.00, and standard deviation 0.00. The count of numerical predictors with outliers is 0 with the minimum percentage of 0.00%, maximum percentage of 0.00%, average percentage of 0.00%,
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=1), 'learning_rate': 0.01, 'n_estimators': 150} |
1016-16-1-4-classification.csv | A multivariate classification time-series dataset consists of 7764 samples and 4 features with 4 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 2 classes with entropy value 0.98 showing a Balanced dataset. Among the 7764 samples the target ground-truth class has changed 301 times representing a percentage of 3.89%. There are 4 features in the dataset
Among the numerical predictors, the series has 4 numerical features detected as Stationary out of the 4 numerical features using the dickey-fuller test and the rest are Unstationary. 4 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 2 seasonality components detected in the numerical predictors. The top 2 common seasonality components are represented using sinusoidal waves. of periods 112,369 The numerical predictors also exhibit skewness values ranging from 0.276. to 0.759 and kurtosis values of 0.18 to 0.62. The fractal dimension analysis yields values ranging from -0.50 to -0.09 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.09, maximum 1.00, mean 0.41, and standard deviation 0.47. The count of numerical predictors with outliers is 3 with the minimum percentage of 0.00%, maximum percentage of 1.62%, average percentage of 0.73%, and standard deviation percentage of 0.84%.
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 40, 'n_estimators': 250} |
1031-42-1-1-5-classification.csv | A multivariate classification time-series dataset consists of 7382 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.41 showing a Unbalanced dataset. Among the 7382 samples the target ground-truth class has changed 1637 times representing a percentage of 22.28%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 14 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.031. to 11.385 and kurtosis values of 0.23 to 179.01. The fractal dimension analysis yields values ranging from -0.69 to -0.11 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.75, maximum 1.00, mean 0.24, and standard deviation 0.45. The count of numerical predictors with outliers is 16 with the minimum percentage of 5.16%, maximum percentage of 17.92%, average percentage of 11.00%, and standard deviation percentage of 3.11%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 0.1, 'n_estimators': 50} |
3001-34.csv | A multivariate classification time-series dataset consists of 216 samples and 1 features with 1 numerical and 0 categorical features. Each instance has a window length of 3. The dataset has a sampling rate of 480.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00
The target column has 2 classes with entropy value 0.92 showing a Unbalanced dataset. Among the 216 samples the target ground-truth class has changed 136 times representing a percentage of 66.67%. There are 1 features in the dataset
Among the numerical predictors, the series has 1 numerical features detected as Stationary out of the 1 numerical features using the dickey-fuller test and the rest are Unstationary. 1 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.045. to 0.045 and kurtosis values of 1.39 to 1.39. The fractal dimension analysis yields values ranging from -1.47 to -1.47 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 1.00, maximum 1.00, mean 1.00, and standard deviation 0.00. The count of numerical predictors with outliers is 0 with the minimum percentage of 0.00%, maximum percentage of 0.00%, average percentage of 0.00%,
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=3), 'learning_rate': 1.0, 'n_estimators': 150} |
1016-10-2-3-classification.csv | A multivariate classification time-series dataset consists of 6408 samples and 6 features with 5 numerical and 1 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0
The target column has 2 classes with entropy value 0.99 showing a Balanced dataset. Among the 6408 samples the target ground-truth class has changed 272 times representing a percentage of 4.27%. There are 6 features in the dataset with a ratio of numerical to categorical features of 5.0.
Among the numerical predictors, the series has 5 numerical features detected as Stationary out of the 5 numerical features using the dickey-fuller test and the rest are Unstationary. 5 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.129. to 0.532 and kurtosis values of 0.30 to 0.83. The fractal dimension analysis yields values ranging from -0.43 to -0.09 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.20, maximum 1.00, mean 0.48, and standard deviation 0.46. The count of numerical predictors with outliers is 2 with the minimum percentage of 0.00%, maximum percentage of 0.42%, average percentage of 0.12%, and standard deviation percentage of 0.19%.
Among the categorical predictors, the count of symbols ranges from 40 to 40 with a minimum entropy value 1.0366372958296777, maximum entropy 1.0366372958296777, mean 1.0366372958296777, and standard deviation 0.0,
The dataset is converted into a simple classification task by extracting the previously described features. | ElasticNetClassifier | {'C': 1000.0, 'l1_ratio': 0.00039999999999999996, 'penalty': 'elasticnet', 'solver': 'saga'} |
1031-36-1-1-6-classification.csv | A multivariate classification time-series dataset consists of 7240 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.43 showing a Unbalanced dataset. Among the 7240 samples the target ground-truth class has changed 432 times representing a percentage of 6.00%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 16 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.070. to 37.852 and kurtosis values of 0.03 to 1433.20. The fractal dimension analysis yields values ranging from -0.87 to -0.09 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.14, maximum 1.00, mean 0.73, and standard deviation 0.39. The count of numerical predictors with outliers is 16 with the minimum percentage of 0.07%, maximum percentage of 33.22%, average percentage of 7.74%, and standard deviation percentage of 13.30%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=1), 'learning_rate': 0.1, 'n_estimators': 50} |
3001-65.csv | A multivariate classification time-series dataset consists of 576 samples and 1 features with 1 numerical and 0 categorical features. Each instance has a window length of 3. The dataset has a sampling rate of 480.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00
The target column has 3 classes with entropy value 1.47 showing a Unbalanced dataset. Among the 576 samples the target ground-truth class has changed 197 times representing a percentage of 34.62%. There are 1 features in the dataset
Among the numerical predictors, the series has 0 numerical features detected as Stationary out of the 1 numerical features using the dickey-fuller test and the rest are Unstationary. 1 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 1.670. to 1.670 and kurtosis values of 1.79 to 1.79. The fractal dimension analysis yields values ranging from -1.02 to -1.02 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 1.00, maximum 1.00, mean 1.00, and standard deviation 0.00. The count of numerical predictors with outliers is 1 with the minimum percentage of 9.49%, maximum percentage of 9.49%, average percentage of 9.49%,
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 10, 'n_estimators': 100} |
1031-29-1-1-2-classification.csv | A multivariate classification time-series dataset consists of 7460 samples and 15 features with 15 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.46 showing a Unbalanced dataset. Among the 7460 samples the target ground-truth class has changed 1367 times representing a percentage of 18.41%. There are 15 features in the dataset
Among the numerical predictors, the series has 15 numerical features detected as Stationary out of the 15 numerical features using the dickey-fuller test and the rest are Unstationary. 13 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.009. to 1.179 and kurtosis values of 0.06 to 3.22. The fractal dimension analysis yields values ranging from -0.59 to -0.11 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.94, maximum 1.00, mean 0.16, and standard deviation 0.51. The count of numerical predictors with outliers is 15 with the minimum percentage of 0.18%, maximum percentage of 11.23%, average percentage of 5.60%, and standard deviation percentage of 3.42%.
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 5, 'n_estimators': 30} |
1031-42-1-1-3-classification.csv | A multivariate classification time-series dataset consists of 7260 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.35 showing a Unbalanced dataset. Among the 7260 samples the target ground-truth class has changed 246 times representing a percentage of 3.40%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 16 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.318. to 12.480 and kurtosis values of 6.21 to 196.14. The fractal dimension analysis yields values ranging from -0.71 to -0.14 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.96, maximum 1.00, mean 0.13, and standard deviation 0.49. The count of numerical predictors with outliers is 16 with the minimum percentage of 22.68%, maximum percentage of 22.68%, average percentage of 22.68%, and standard deviation percentage of 0.00%.
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 10, 'n_estimators': 50} |
1021-2-classification.csv | A multivariate classification time-series dataset consists of 6996 samples and 7 features with 4 numerical and 3 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0 and standard deviation 0.0.
The target column has 3 classes with entropy value 1.46 showing a Unbalanced dataset. Among the 6996 samples the target ground-truth class has changed 702 times representing a percentage of 10.08%. There are 7 features in the dataset with a ratio of numerical to categorical features of 1.3333333333333333.
Among the numerical predictors, the series has 4 numerical features detected as Stationary out of the 4 numerical features using the dickey-fuller test and the rest are Unstationary. 3 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 46 seasonality components detected in the numerical predictors. The top 3 common seasonality components are represented using sinusoidal waves. of periods 466,538,1166 The numerical predictors also exhibit skewness values ranging from 0.309. to 3.853 and kurtosis values of 0.57 to 18.48. The fractal dimension analysis yields values ranging from -0.35 to -0.15 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.84, maximum 1.00, mean 0.11, and standard deviation 0.72. The count of numerical predictors with outliers is 1 with the minimum percentage of 0.00%, maximum percentage of 10.33%, average percentage of 2.58%, and standard deviation percentage of 5.16%.
Among the categorical predictors, the count of symbols ranges from 5 to 28 with a minimum entropy value 0.0895126156217436, maximum entropy 2.1071047601634865, mean 0.8507667872900294, and standard deviation 0.8949862170599643,
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
1020-1-1-classification.csv | A multivariate classification time-series dataset consists of 7008 samples and 11 features with 9 numerical and 2 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0 and standard deviation 0.0.
The target column has 3 classes with entropy value 1.46 showing a Unbalanced dataset. Among the 7008 samples the target ground-truth class has changed 663 times representing a percentage of 9.51%. There are 11 features in the dataset with a ratio of numerical to categorical features of 4.5.
Among the numerical predictors, the series has 9 numerical features detected as Stationary out of the 9 numerical features using the dickey-fuller test and the rest are Unstationary. 8 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 19 seasonality components detected in the numerical predictors. The top 3 common seasonality components are represented using sinusoidal waves. of periods 219,226,1401 The numerical predictors also exhibit skewness values ranging from 0.004. to 2.388 and kurtosis values of 0.52 to 10.26. The fractal dimension analysis yields values ranging from -0.71 to -0.14 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.73, maximum 1.00, mean 0.14, and standard deviation 0.46. The count of numerical predictors with outliers is 6 with the minimum percentage of 0.00%, maximum percentage of 9.58%, average percentage of 3.42%, and standard deviation percentage of 3.19%.
Among the categorical predictors, the count of symbols ranges from 17 to 64 with a minimum entropy value 0.44837030789530447, maximum entropy 3.8153891750864126, mean 2.1318797414908586, and standard deviation 1.683509433595554,
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 10, 'n_estimators': 30} |
1031-60-1-1-5-classification.csv | A multivariate classification time-series dataset consists of 6576 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.51 showing a Unbalanced dataset. Among the 6576 samples the target ground-truth class has changed 1381 times representing a percentage of 21.11%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 14 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.002. to 11.744 and kurtosis values of 0.02 to 176.81. The fractal dimension analysis yields values ranging from -0.71 to -0.13 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.97, maximum 1.00, mean 0.12, and standard deviation 0.47. The count of numerical predictors with outliers is 16 with the minimum percentage of 0.57%, maximum percentage of 6.80%, average percentage of 4.12%, and standard deviation percentage of 1.77%.
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 10, 'n_estimators': 30} |
1031-29-1-1-6-classification.csv | A multivariate classification time-series dataset consists of 7464 samples and 15 features with 15 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.49 showing a Unbalanced dataset. Among the 7464 samples the target ground-truth class has changed 1660 times representing a percentage of 22.34%. There are 15 features in the dataset
Among the numerical predictors, the series has 15 numerical features detected as Stationary out of the 15 numerical features using the dickey-fuller test and the rest are Unstationary. 15 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.066. to 1.549 and kurtosis values of 0.08 to 6.47. The fractal dimension analysis yields values ranging from -0.61 to -0.12 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.96, maximum 1.00, mean 0.13, and standard deviation 0.51. The count of numerical predictors with outliers is 15 with the minimum percentage of 0.20%, maximum percentage of 14.51%, average percentage of 4.67%, and standard deviation percentage of 4.29%.
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
1031-32-2-1-1-classification.csv | A multivariate classification time-series dataset consists of 7324 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.44 showing a Unbalanced dataset. Among the 7324 samples the target ground-truth class has changed 1428 times representing a percentage of 19.59%. There are 16 features in the dataset
Among the numerical predictors, the series has 13 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 14 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.042. to 16.369 and kurtosis values of 0.03 to 399.74. The fractal dimension analysis yields values ranging from -0.62 to -0.11 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.90, maximum 1.00, mean 0.11, and standard deviation 0.51. The count of numerical predictors with outliers is 16 with the minimum percentage of 0.34%, maximum percentage of 13.62%, average percentage of 7.54%, and standard deviation percentage of 4.07%.
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 5, 'n_estimators': 50} |
3001-28.csv | A multivariate classification time-series dataset consists of 768 samples and 2 features with 2 numerical and 0 categorical features. Each instance has a window length of 4. The dataset has a sampling rate of 360.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 4 classes with entropy value 1.75 showing a Unbalanced dataset. Among the 768 samples the target ground-truth class has changed 751 times representing a percentage of 99.60%. There are 2 features in the dataset
Among the numerical predictors, the series has 2 numerical features detected as Stationary out of the 2 numerical features using the dickey-fuller test and the rest are Unstationary. 1 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 1 seasonality components detected in the numerical predictors. The top 1 common seasonality components are represented using sinusoidal waves. of periods 27 The numerical predictors also exhibit skewness values ranging from 0.065. to 1.784 and kurtosis values of 0.53 to 2.64. The fractal dimension analysis yields values ranging from -1.20 to -1.07 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.51, maximum 1.00, mean 0.24, and standard deviation 0.76. The count of numerical predictors with outliers is 1 with the minimum percentage of 0.00%, maximum percentage of 10.74%, average percentage of 5.37%, and standard deviation percentage of 7.60%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=5), 'learning_rate': 1.0, 'n_estimators': 150} |
1020-11-1-classification.csv | A multivariate classification time-series dataset consists of 7008 samples and 11 features with 9 numerical and 2 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0 and standard deviation 0.0.
The target column has 3 classes with entropy value 1.44 showing a Unbalanced dataset. Among the 7008 samples the target ground-truth class has changed 809 times representing a percentage of 11.60%. There are 11 features in the dataset with a ratio of numerical to categorical features of 4.5.
Among the numerical predictors, the series has 8 numerical features detected as Stationary out of the 9 numerical features using the dickey-fuller test and the rest are Unstationary. 8 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 30 seasonality components detected in the numerical predictors. The top 3 common seasonality components are represented using sinusoidal waves. of periods 1401,1752,2336 The numerical predictors also exhibit skewness values ranging from 0.050. to 2.183 and kurtosis values of 0.69 to 5.85. The fractal dimension analysis yields values ranging from -0.75 to -0.14 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.76, maximum 1.00, mean 0.16, and standard deviation 0.47. The count of numerical predictors with outliers is 6 with the minimum percentage of 0.00%, maximum percentage of 9.19%, average percentage of 3.34%, and standard deviation percentage of 3.06%.
Among the categorical predictors, the count of symbols ranges from 17 to 63 with a minimum entropy value 0.38849685263596306, maximum entropy 3.890878388180946, mean 2.1396876204084543, and standard deviation 1.7511907677724914,
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 1.0, 'n_estimators': 50} |
2007-41.csv | A multivariate classification time-series dataset consists of 1262 samples and 9 features with 8 numerical and 1 categorical features. Each instance has a window length of 7. The dataset has a sampling rate of 1440.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 29 with mean 12.50 and standard deviation 13.03. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0
The target column has 3 classes with entropy value 0.61 showing a Unbalanced dataset. Among the 1262 samples the target ground-truth class has changed 152 times representing a percentage of 12.21%. There are 9 features in the dataset with a ratio of numerical to categorical features of 8.0.
Among the numerical predictors, the series has 6 numerical features detected as Stationary out of the 8 numerical features using the dickey-fuller test and the rest are Unstationary. 7 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 19 seasonality components detected in the numerical predictors. The top 3 common seasonality components are represented using sinusoidal waves. of periods 114,210,420 The numerical predictors also exhibit skewness values ranging from 0.062. to 9.427 and kurtosis values of 0.28 to 103.88. The fractal dimension analysis yields values ranging from -0.91 to -0.57 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.70, maximum 1.00, mean 0.16, and standard deviation 0.44. The count of numerical predictors with outliers is 3 with the minimum percentage of 0.00%, maximum percentage of 15.18%, average percentage of 2.62%, and standard deviation percentage of 5.32%.
Among the categorical predictors, the count of symbols ranges from 3 to 3 with a minimum entropy value 0.6069739075939655, maximum entropy 0.6069739075939655, mean 0.6069739075939655, and standard deviation 0.0,
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=3), 'learning_rate': 0.01, 'n_estimators': 50} |
1031-33-2-1-2-classification.csv | A multivariate classification time-series dataset consists of 7172 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.37 showing a Unbalanced dataset. Among the 7172 samples the target ground-truth class has changed 837 times representing a percentage of 11.73%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 16 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.134. to 21.123 and kurtosis values of 1.26 to 640.93. The fractal dimension analysis yields values ranging from -0.67 to -0.12 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.85, maximum 1.00, mean 0.07, and standard deviation 0.48. The count of numerical predictors with outliers is 16 with the minimum percentage of 2.65%, maximum percentage of 42.55%, average percentage of 37.55%, and standard deviation percentage of 10.52%.
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
1020-32-1-classification.csv | A multivariate classification time-series dataset consists of 6992 samples and 11 features with 9 numerical and 2 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0 and standard deviation 0.0.
The target column has 3 classes with entropy value 1.45 showing a Unbalanced dataset. Among the 6992 samples the target ground-truth class has changed 1144 times representing a percentage of 16.44%. There are 11 features in the dataset with a ratio of numerical to categorical features of 4.5.
Among the numerical predictors, the series has 9 numerical features detected as Stationary out of the 9 numerical features using the dickey-fuller test and the rest are Unstationary. 9 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 56 seasonality components detected in the numerical predictors. The top 3 common seasonality components are represented using sinusoidal waves. of periods 1748,2330,3496 The numerical predictors also exhibit skewness values ranging from 0.028. to 3.155 and kurtosis values of 0.62 to 14.74. The fractal dimension analysis yields values ranging from -0.75 to -0.11 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.74, maximum 1.00, mean 0.13, and standard deviation 0.48. The count of numerical predictors with outliers is 6 with the minimum percentage of 0.00%, maximum percentage of 10.64%, average percentage of 3.53%, and standard deviation percentage of 3.50%.
Among the categorical predictors, the count of symbols ranges from 17 to 57 with a minimum entropy value 0.4009512591270654, maximum entropy 3.890763316267035, mean 2.14585728769705, and standard deviation 1.7449060285699847,
The dataset is converted into a simple classification task by extracting the previously described features. | LassoClassifier | {'C': 223.60679774997894, 'penalty': 'l1', 'solver': 'saga'} |
1016-11-3-2-classification.csv | A multivariate classification time-series dataset consists of 7109 samples and 12 features with 4 numerical and 8 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0 and standard deviation 0.0.
The target column has 2 classes with entropy value 0.98 showing a Balanced dataset. Among the 7109 samples the target ground-truth class has changed 306 times representing a percentage of 4.33%. There are 12 features in the dataset with a ratio of numerical to categorical features of 0.5.
Among the numerical predictors, the series has 4 numerical features detected as Stationary out of the 4 numerical features using the dickey-fuller test and the rest are Unstationary. 4 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 1 seasonality components detected in the numerical predictors. The top 1 common seasonality components are represented using sinusoidal waves. of periods 103 The numerical predictors also exhibit skewness values ranging from 0.014. to 0.640 and kurtosis values of 0.10 to 0.46. The fractal dimension analysis yields values ranging from -0.45 to -0.10 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.08, maximum 1.00, mean 0.59, and standard deviation 0.50. The count of numerical predictors with outliers is 3 with the minimum percentage of 0.00%, maximum percentage of 1.55%, average percentage of 0.69%, and standard deviation percentage of 0.75%.
Among the categorical predictors, the count of symbols ranges from 9 to 68 with a minimum entropy value 1.1526627332728498, maximum entropy 5.46373859395374, mean 3.708604491796955, and standard deviation 1.514849483180694,
The dataset is converted into a simple classification task by extracting the previously described features. | ElasticNetClassifier | {'C': 1000.0, 'l1_ratio': 0.0001, 'penalty': 'elasticnet', 'solver': 'saga'} |
3001-21.csv | A multivariate classification time-series dataset consists of 168 samples and 1 features with 1 numerical and 0 categorical features. Each instance has a window length of 7. The dataset has a sampling rate of 1440.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00
The target column has 2 classes with entropy value 1.00 showing a Balanced dataset. Among the 168 samples the target ground-truth class has changed 1 times representing a percentage of 0.63%. There are 1 features in the dataset
Among the numerical predictors, the series has 0 numerical features detected as Stationary out of the 1 numerical features using the dickey-fuller test and the rest are Unstationary. 0 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.079. to 0.079 and kurtosis values of 1.96 to 1.96. The fractal dimension analysis yields values ranging from -0.86 to -0.86 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 1.00, maximum 1.00, mean 1.00, and standard deviation 0.00. The count of numerical predictors with outliers is 0 with the minimum percentage of 0.00%, maximum percentage of 0.00%, average percentage of 0.00%,
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 10, 'n_estimators': 100} |
1030-324-classification.csv | A multivariate classification time-series dataset consists of 4140 samples and 5 features with 5 numerical and 0 categorical features. Each instance has a window length of 7. The dataset has a sampling rate of 1440.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 2 classes with entropy value 0.97 showing a Unbalanced dataset. Among the 4140 samples the target ground-truth class has changed 22 times representing a percentage of 0.53%. There are 5 features in the dataset
Among the numerical predictors, the series has 1 numerical features detected as Stationary out of the 5 numerical features using the dickey-fuller test and the rest are Unstationary. 5 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.545. to 5.771 and kurtosis values of 0.76 to 65.75. The fractal dimension analysis yields values ranging from -0.62 to -0.29 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 0.06, maximum 1.00, mean 0.70, and standard deviation 0.43. The count of numerical predictors with outliers is 1 with the minimum percentage of 0.00%, maximum percentage of 5.75%, average percentage of 1.15%, and standard deviation percentage of 2.57%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 0.1, 'n_estimators': 50} |
1016-2-3-3-classification.csv | A multivariate classification time-series dataset consists of 7385 samples and 8 features with 4 numerical and 4 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0 and standard deviation 0.0.
The target column has 2 classes with entropy value 0.97 showing a Unbalanced dataset. Among the 7385 samples the target ground-truth class has changed 252 times representing a percentage of 3.43%. There are 8 features in the dataset with a ratio of numerical to categorical features of 1.0.
Among the numerical predictors, the series has 4 numerical features detected as Stationary out of the 4 numerical features using the dickey-fuller test and the rest are Unstationary. 3 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.039. to 0.545 and kurtosis values of 0.03 to 0.71. The fractal dimension analysis yields values ranging from -0.44 to -0.08 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.03, maximum 1.00, mean 0.61, and standard deviation 0.48. The count of numerical predictors with outliers is 3 with the minimum percentage of 0.00%, maximum percentage of 0.76%, average percentage of 0.36%, and standard deviation percentage of 0.32%.
Among the categorical predictors, the count of symbols ranges from 36 to 62 with a minimum entropy value 1.5557503447174752, maximum entropy 5.233353299634948, mean 4.123066341401343, and standard deviation 1.4904130683078063,
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 10, 'n_estimators': 50} |
1031-57-1-1-1-classification.csv | A multivariate classification time-series dataset consists of 7456 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.45 showing a Unbalanced dataset. Among the 7456 samples the target ground-truth class has changed 1469 times representing a percentage of 19.79%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 13 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.016. to 13.851 and kurtosis values of 0.14 to 242.21. The fractal dimension analysis yields values ranging from -0.63 to -0.11 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.95, maximum 1.00, mean 0.13, and standard deviation 0.53. The count of numerical predictors with outliers is 16 with the minimum percentage of 0.11%, maximum percentage of 13.31%, average percentage of 5.21%, and standard deviation percentage of 4.42%.
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 40, 'n_estimators': 200} |
1016-9-2-4-classification.csv | A multivariate classification time-series dataset consists of 7109 samples and 12 features with 4 numerical and 8 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0 and standard deviation 0.0.
The target column has 2 classes with entropy value 1.00 showing a Balanced dataset. Among the 7109 samples the target ground-truth class has changed 337 times representing a percentage of 4.76%. There are 12 features in the dataset with a ratio of numerical to categorical features of 0.5.
Among the numerical predictors, the series has 4 numerical features detected as Stationary out of the 4 numerical features using the dickey-fuller test and the rest are Unstationary. 4 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 4 seasonality components detected in the numerical predictors. The top 3 common seasonality components are represented using sinusoidal waves. of periods 197,1184,1184 The numerical predictors also exhibit skewness values ranging from 0.117. to 0.354 and kurtosis values of 0.05 to 0.37. The fractal dimension analysis yields values ranging from -0.42 to -0.08 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.09, maximum 1.00, mean 0.58, and standard deviation 0.51. The count of numerical predictors with outliers is 4 with the minimum percentage of 0.08%, maximum percentage of 0.78%, average percentage of 0.42%, and standard deviation percentage of 0.29%.
Among the categorical predictors, the count of symbols ranges from 9 to 64 with a minimum entropy value 1.2029113656607329, maximum entropy 5.048299320319803, mean 3.5798824355296923, and standard deviation 1.3887038372258869,
The dataset is converted into a simple classification task by extracting the previously described features. | LassoClassifier | {'C': 4728.708045015879, 'penalty': 'l1', 'solver': 'saga'} |
1031-25-2-1-2-classification.csv | A multivariate classification time-series dataset consists of 7380 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.37 showing a Unbalanced dataset. Among the 7380 samples the target ground-truth class has changed 921 times representing a percentage of 12.54%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 15 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.036. to 20.922 and kurtosis values of 1.18 to 522.55. The fractal dimension analysis yields values ranging from -0.69 to -0.12 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.96, maximum 1.00, mean 0.15, and standard deviation 0.49. The count of numerical predictors with outliers is 16 with the minimum percentage of 2.45%, maximum percentage of 45.64%, average percentage of 40.53%, and standard deviation percentage of 13.05%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 1.0, 'n_estimators': 50} |
1016-11-3-3-classification.csv | A multivariate classification time-series dataset consists of 7109 samples and 12 features with 4 numerical and 8 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0 and standard deviation 0.0.
The target column has 2 classes with entropy value 1.00 showing a Balanced dataset. Among the 7109 samples the target ground-truth class has changed 394 times representing a percentage of 5.57%. There are 12 features in the dataset with a ratio of numerical to categorical features of 0.5.
Among the numerical predictors, the series has 4 numerical features detected as Stationary out of the 4 numerical features using the dickey-fuller test and the rest are Unstationary. 3 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 7 seasonality components detected in the numerical predictors. The top 3 common seasonality components are represented using sinusoidal waves. of periods 323,374,789 The numerical predictors also exhibit skewness values ranging from 0.095. to 0.821 and kurtosis values of 0.23 to 0.72. The fractal dimension analysis yields values ranging from -0.41 to -0.10 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 0.07, maximum 1.00, mean 0.65, and standard deviation 0.42. The count of numerical predictors with outliers is 3 with the minimum percentage of 0.00%, maximum percentage of 2.36%, average percentage of 0.70%, and standard deviation percentage of 1.11%.
Among the categorical predictors, the count of symbols ranges from 9 to 70 with a minimum entropy value 1.5210808411835828, maximum entropy 5.347519588572782, mean 3.733769867190697, and standard deviation 1.4565790018371942,
The dataset is converted into a simple classification task by extracting the previously described features. | ElasticNetClassifier | {'C': 1000.0, 'l1_ratio': 0.00055, 'penalty': 'elasticnet', 'solver': 'saga'} |
1031-50-1-1-2-classification.csv | A multivariate classification time-series dataset consists of 6768 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.36 showing a Unbalanced dataset. Among the 6768 samples the target ground-truth class has changed 1269 times representing a percentage of 18.84%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 12 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.010. to 21.337 and kurtosis values of 0.20 to 666.13. The fractal dimension analysis yields values ranging from -0.60 to -0.11 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.89, maximum 1.00, mean 0.09, and standard deviation 0.49. The count of numerical predictors with outliers is 16 with the minimum percentage of 3.88%, maximum percentage of 18.65%, average percentage of 13.04%, and standard deviation percentage of 4.48%.
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 10, 'n_estimators': 30} |
1016-22-2-1-classification.csv | A multivariate classification time-series dataset consists of 7107 samples and 8 features with 8 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.42 showing a Unbalanced dataset. Among the 7107 samples the target ground-truth class has changed 700 times representing a percentage of 9.90%. There are 8 features in the dataset
Among the numerical predictors, the series has 8 numerical features detected as Stationary out of the 8 numerical features using the dickey-fuller test and the rest are Unstationary. 7 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 31 seasonality components detected in the numerical predictors. The top 3 common seasonality components are represented using sinusoidal waves. of periods 1015,1184,1776 The numerical predictors also exhibit skewness values ranging from 0.188. to 10.188 and kurtosis values of 0.08 to 142.30. The fractal dimension analysis yields values ranging from -0.58 to -0.07 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.44, maximum 1.00, mean 0.18, and standard deviation 0.45. The count of numerical predictors with outliers is 7 with the minimum percentage of 0.00%, maximum percentage of 6.87%, average percentage of 2.34%, and standard deviation percentage of 2.22%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 1.0, 'n_estimators': 50} |
1016-2-2-5-classification.csv | A multivariate classification time-series dataset consists of 7385 samples and 8 features with 4 numerical and 4 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0 and standard deviation 0.0.
The target column has 2 classes with entropy value 1.00 showing a Balanced dataset. Among the 7385 samples the target ground-truth class has changed 278 times representing a percentage of 3.78%. There are 8 features in the dataset with a ratio of numerical to categorical features of 1.0.
Among the numerical predictors, the series has 4 numerical features detected as Stationary out of the 4 numerical features using the dickey-fuller test and the rest are Unstationary. 2 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.085. to 0.857 and kurtosis values of 0.44 to 1.23. The fractal dimension analysis yields values ranging from -0.48 to -0.08 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 0.26, maximum 1.00, mean 0.72, and standard deviation 0.34. The count of numerical predictors with outliers is 2 with the minimum percentage of 0.00%, maximum percentage of 1.52%, average percentage of 0.44%, and standard deviation percentage of 0.73%.
Among the categorical predictors, the count of symbols ranges from 36 to 59 with a minimum entropy value 1.3726660361760685, maximum entropy 5.261536331466376, mean 4.098124791222294, and standard deviation 1.5839224241577705,
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 0.1, 'n_estimators': 50} |
1030-306-classification.csv | A multivariate classification time-series dataset consists of 4140 samples and 5 features with 5 numerical and 0 categorical features. Each instance has a window length of 7. The dataset has a sampling rate of 1440.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 4 classes with entropy value 1.90 showing a Unbalanced dataset. Among the 4140 samples the target ground-truth class has changed 23 times representing a percentage of 0.56%. There are 5 features in the dataset
Among the numerical predictors, the series has 1 numerical features detected as Stationary out of the 5 numerical features using the dickey-fuller test and the rest are Unstationary. 5 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.975. to 6.164 and kurtosis values of 0.23 to 65.12. The fractal dimension analysis yields values ranging from -0.61 to -0.34 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.37, maximum 1.00, mean 0.56, and standard deviation 0.64. The count of numerical predictors with outliers is 1 with the minimum percentage of 0.00%, maximum percentage of 6.06%, average percentage of 1.21%, and standard deviation percentage of 2.71%.
The dataset is converted into a simple classification task by extracting the previously described features. | ElasticNetClassifier | {'C': 1000.0, 'l1_ratio': 0.0001, 'penalty': 'elasticnet', 'solver': 'saga'} |
1031-15-1-1-1-classification.csv | A multivariate classification time-series dataset consists of 7590 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.46 showing a Unbalanced dataset. Among the 7590 samples the target ground-truth class has changed 1510 times representing a percentage of 19.98%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 13 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.005. to 12.588 and kurtosis values of 0.01 to 197.92. The fractal dimension analysis yields values ranging from -0.60 to -0.11 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.94, maximum 1.00, mean 0.10, and standard deviation 0.50. The count of numerical predictors with outliers is 16 with the minimum percentage of 0.93%, maximum percentage of 12.19%, average percentage of 5.93%, and standard deviation percentage of 3.55%.
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 10, 'n_estimators': 30} |
1031-9-1-1-5-classification.csv | A multivariate classification time-series dataset consists of 7488 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.35 showing a Unbalanced dataset. Among the 7488 samples the target ground-truth class has changed 209 times representing a percentage of 2.80%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 16 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 1.859. to 38.323 and kurtosis values of 9.37 to 1473.65. The fractal dimension analysis yields values ranging from -0.46 to -0.07 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 0.03, maximum 1.00, mean 0.85, and standard deviation 0.29. The count of numerical predictors with outliers is 16 with the minimum percentage of 12.80%, maximum percentage of 12.80%, average percentage of 12.80%, and standard deviation percentage of 0.00%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 0.1, 'n_estimators': 50} |
1030-260-classification.csv | A multivariate classification time-series dataset consists of 4140 samples and 5 features with 5 numerical and 0 categorical features. Each instance has a window length of 7. The dataset has a sampling rate of 1440.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 2 classes with entropy value 1.00 showing a Balanced dataset. Among the 4140 samples the target ground-truth class has changed 25 times representing a percentage of 0.61%. There are 5 features in the dataset
Among the numerical predictors, the series has 1 numerical features detected as Stationary out of the 5 numerical features using the dickey-fuller test and the rest are Unstationary. 1 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.065. to 3.180 and kurtosis values of 0.73 to 29.18. The fractal dimension analysis yields values ranging from -0.64 to -0.33 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.53, maximum 1.00, mean 0.50, and standard deviation 0.69. The count of numerical predictors with outliers is 1 with the minimum percentage of 0.00%, maximum percentage of 4.61%, average percentage of 0.92%, and standard deviation percentage of 2.06%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 1.0, 'n_estimators': 50} |
1020-69-4-classification.csv | A multivariate classification time-series dataset consists of 7012 samples and 11 features with 9 numerical and 2 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0 and standard deviation 0.0.
The target column has 3 classes with entropy value 1.35 showing a Unbalanced dataset. Among the 7012 samples the target ground-truth class has changed 627 times representing a percentage of 8.99%. There are 11 features in the dataset with a ratio of numerical to categorical features of 4.5.
Among the numerical predictors, the series has 8 numerical features detected as Stationary out of the 9 numerical features using the dickey-fuller test and the rest are Unstationary. 9 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 36 seasonality components detected in the numerical predictors. The top 3 common seasonality components are represented using sinusoidal waves. of periods 779,1168,1753 The numerical predictors also exhibit skewness values ranging from 0.359. to 2.028 and kurtosis values of 0.44 to 6.37. The fractal dimension analysis yields values ranging from -0.67 to -0.14 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.79, maximum 1.00, mean 0.14, and standard deviation 0.52. The count of numerical predictors with outliers is 7 with the minimum percentage of 0.00%, maximum percentage of 8.40%, average percentage of 2.93%, and standard deviation percentage of 2.80%.
Among the categorical predictors, the count of symbols ranges from 17 to 67 with a minimum entropy value 0.4687768858635983, maximum entropy 3.929813113804487, mean 2.1992949998340428, and standard deviation 1.7305181139704444,
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 1.0, 'n_estimators': 50} |
1031-51-1-1-3-classification.csv | A multivariate classification time-series dataset consists of 6720 samples and 16 features with 15 numerical and 1 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0
The target column has 3 classes with entropy value 1.37 showing a Unbalanced dataset. Among the 6720 samples the target ground-truth class has changed 921 times representing a percentage of 13.78%. There are 16 features in the dataset with a ratio of numerical to categorical features of 15.0.
Among the numerical predictors, the series has 15 numerical features detected as Stationary out of the 15 numerical features using the dickey-fuller test and the rest are Unstationary. 13 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.050. to 1.847 and kurtosis values of 0.98 to 10.97. The fractal dimension analysis yields values ranging from -0.63 to -0.11 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.91, maximum 1.00, mean 0.12, and standard deviation 0.50. The count of numerical predictors with outliers is 15 with the minimum percentage of 37.53%, maximum percentage of 45.11%, average percentage of 43.35%, and standard deviation percentage of 2.88%.
Among the categorical predictors, the count of symbols ranges from 38 to 38 with a minimum entropy value 0.12266657826549437, maximum entropy 0.12266657826549437, mean 0.12266657826549437, and standard deviation 0.0,
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
1034-3-4-classification.csv | A multivariate classification time-series dataset consists of 7963 samples and 6 features with 6 numerical and 0 categorical features. Each instance has a window length of 4. The dataset has a sampling rate of 15.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 2 classes with entropy value 0.97 showing a Unbalanced dataset. Among the 7963 samples the target ground-truth class has changed 221 times representing a percentage of 2.78%. There are 6 features in the dataset
Among the numerical predictors, the series has 4 numerical features detected as Stationary out of the 6 numerical features using the dickey-fuller test and the rest are Unstationary. 6 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.097. to 1.794 and kurtosis values of 0.25 to 2.53. The fractal dimension analysis yields values ranging from -0.73 to -0.44 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.36, maximum 1.00, mean 0.23, and standard deviation 0.50. The count of numerical predictors with outliers is 5 with the minimum percentage of 0.00%, maximum percentage of 10.11%, average percentage of 3.81%, and standard deviation percentage of 3.38%.
The dataset is converted into a simple classification task by extracting the previously described features. | ElasticNetClassifier | {'C': 1000.0, 'l1_ratio': 0.0001, 'penalty': 'elasticnet', 'solver': 'saga'} |
3001-95.csv | A multivariate classification time-series dataset consists of 288 samples and 1 features with 1 numerical and 0 categorical features. Each instance has a window length of 4. The dataset has a sampling rate of 360.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00
The target column has 3 classes with entropy value 1.50 showing a Unbalanced dataset. Among the 288 samples the target ground-truth class has changed 274 times representing a percentage of 99.64%. There are 1 features in the dataset
Among the numerical predictors, the series has 1 numerical features detected as Stationary out of the 1 numerical features using the dickey-fuller test and the rest are Unstationary. 1 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.927. to 0.927 and kurtosis values of 0.09 to 0.09. The fractal dimension analysis yields values ranging from -1.20 to -1.20 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 1.00, maximum 1.00, mean 1.00, and standard deviation 0.00. The count of numerical predictors with outliers is 0 with the minimum percentage of 0.00%, maximum percentage of 0.00%, average percentage of 0.00%,
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=3), 'learning_rate': 0.5, 'n_estimators': 50} |
1030-14-classification.csv | A multivariate classification time-series dataset consists of 4140 samples and 5 features with 5 numerical and 0 categorical features. Each instance has a window length of 7. The dataset has a sampling rate of 1440.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 2 classes with entropy value 0.97 showing a Unbalanced dataset. Among the 4140 samples the target ground-truth class has changed 35 times representing a percentage of 0.85%. There are 5 features in the dataset
Among the numerical predictors, the series has 1 numerical features detected as Stationary out of the 5 numerical features using the dickey-fuller test and the rest are Unstationary. 5 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 1.391. to 7.925 and kurtosis values of 1.83 to 159.41. The fractal dimension analysis yields values ranging from -0.65 to -0.32 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 0.15, maximum 1.00, mean 0.73, and standard deviation 0.39. The count of numerical predictors with outliers is 5 with the minimum percentage of 3.90%, maximum percentage of 6.14%, average percentage of 5.24%, and standard deviation percentage of 1.14%.
The dataset is converted into a simple classification task by extracting the previously described features. | ElasticNetClassifier | {'C': 1000.0, 'l1_ratio': 0.0001, 'penalty': 'elasticnet', 'solver': 'saga'} |
1031-50-2-1-4-classification.csv | A multivariate classification time-series dataset consists of 7044 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 2 classes with entropy value 0.97 showing a Unbalanced dataset. Among the 7044 samples the target ground-truth class has changed 679 times representing a percentage of 9.69%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 15 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.006. to 14.005 and kurtosis values of 0.07 to 251.58. The fractal dimension analysis yields values ranging from -0.78 to -0.11 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.83, maximum 1.00, mean 0.10, and standard deviation 0.50. The count of numerical predictors with outliers is 15 with the minimum percentage of 0.00%, maximum percentage of 12.94%, average percentage of 4.53%, and standard deviation percentage of 4.52%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 1.0, 'n_estimators': 50} |
1031-4-3-1-4-classification.csv | A multivariate classification time-series dataset consists of 7644 samples and 16 features with 15 numerical and 1 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0
The target column has 3 classes with entropy value 1.38 showing a Unbalanced dataset. Among the 7644 samples the target ground-truth class has changed 1361 times representing a percentage of 17.88%. There are 16 features in the dataset with a ratio of numerical to categorical features of 15.0.
Among the numerical predictors, the series has 13 numerical features detected as Stationary out of the 15 numerical features using the dickey-fuller test and the rest are Unstationary. 13 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.042. to 1.973 and kurtosis values of 0.12 to 6.23. The fractal dimension analysis yields values ranging from -0.63 to -0.12 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.86, maximum 1.00, mean 0.10, and standard deviation 0.58. The count of numerical predictors with outliers is 15 with the minimum percentage of 5.49%, maximum percentage of 16.85%, average percentage of 11.88%, and standard deviation percentage of 3.96%.
Among the categorical predictors, the count of symbols ranges from 94 to 94 with a minimum entropy value 0.4477999259360229, maximum entropy 0.4477999259360229, mean 0.4477999259360229, and standard deviation 0.0,
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 10, 'n_estimators': 30} |
1029-18-classification.csv | A multivariate classification time-series dataset consists of 3503 samples and 4 features with 4 numerical and 0 categorical features. Each instance has a window length of 7. The dataset has a sampling rate of 1440.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.38 showing a Unbalanced dataset. Among the 3503 samples the target ground-truth class has changed 50 times representing a percentage of 1.43%. There are 4 features in the dataset
Among the numerical predictors, the series has 1 numerical features detected as Stationary out of the 4 numerical features using the dickey-fuller test and the rest are Unstationary. 4 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.446. to 2.813 and kurtosis values of 0.41 to 14.73. The fractal dimension analysis yields values ranging from -0.62 to -0.32 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.32, maximum 1.00, mean 0.51, and standard deviation 0.63. The count of numerical predictors with outliers is 1 with the minimum percentage of 0.00%, maximum percentage of 4.99%, average percentage of 1.25%, and standard deviation percentage of 2.50%.
The dataset is converted into a simple classification task by extracting the previously described features. | ElasticNetClassifier | {'C': 1000.0, 'l1_ratio': 0.0001, 'penalty': 'elasticnet', 'solver': 'saga'} |
1030-16-classification.csv | A multivariate classification time-series dataset consists of 4140 samples and 5 features with 5 numerical and 0 categorical features. Each instance has a window length of 7. The dataset has a sampling rate of 1440.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 2 classes with entropy value 0.97 showing a Unbalanced dataset. Among the 4140 samples the target ground-truth class has changed 23 times representing a percentage of 0.56%. There are 5 features in the dataset
Among the numerical predictors, the series has 1 numerical features detected as Stationary out of the 5 numerical features using the dickey-fuller test and the rest are Unstationary. 5 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 1.447. to 9.180 and kurtosis values of 1.59 to 157.20. The fractal dimension analysis yields values ranging from -0.61 to -0.29 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.01, maximum 1.00, mean 0.68, and standard deviation 0.47. The count of numerical predictors with outliers is 5 with the minimum percentage of 3.25%, maximum percentage of 5.70%, average percentage of 3.84%, and standard deviation percentage of 1.04%.
The dataset is converted into a simple classification task by extracting the previously described features. | ElasticNetClassifier | {'C': 1000.0, 'l1_ratio': 0.0001, 'penalty': 'elasticnet', 'solver': 'saga'} |
1031-28-1-1-1-classification.csv | A multivariate classification time-series dataset consists of 7574 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.43 showing a Unbalanced dataset. Among the 7574 samples the target ground-truth class has changed 1574 times representing a percentage of 20.88%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 16 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.138. to 13.460 and kurtosis values of 0.07 to 230.06. The fractal dimension analysis yields values ranging from -0.64 to -0.12 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.96, maximum 1.00, mean 0.11, and standard deviation 0.47. The count of numerical predictors with outliers is 16 with the minimum percentage of 1.76%, maximum percentage of 15.84%, average percentage of 6.91%, and standard deviation percentage of 3.76%.
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.01, 'max_depth': 3, 'n_estimators': 20, 'reg_lambda': 0.2} |
1019-2-classification.csv | A multivariate classification time-series dataset consists of 1100 samples and 17 features with 16 numerical and 1 categorical features. Each instance has a window length of 4. The dataset has a sampling rate of 15.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0
The target column has 2 classes with entropy value 0.99 showing a Balanced dataset. Among the 1100 samples the target ground-truth class has changed 22 times representing a percentage of 2.03%. There are 17 features in the dataset with a ratio of numerical to categorical features of 16.0.
Among the numerical predictors, the series has 14 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 15 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.123. to 4.965 and kurtosis values of 0.11 to 38.34. The fractal dimension analysis yields values ranging from -0.66 to -0.00 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.66, maximum 1.00, mean 0.26, and standard deviation 0.44. The count of numerical predictors with outliers is 7 with the minimum percentage of 0.00%, maximum percentage of 16.02%, average percentage of 3.26%, and standard deviation percentage of 5.73%.
Among the categorical predictors, the count of symbols ranges from 6 to 6 with a minimum entropy value 0.1005636453394341, maximum entropy 0.1005636453394341, mean 0.1005636453394341, and standard deviation 0.0,
The dataset is converted into a simple classification task by extracting the previously described features. | ElasticNetClassifier | {'C': 1000.0, 'l1_ratio': 0.0001, 'penalty': 'elasticnet', 'solver': 'saga'} |
1031-53-1-1-4-classification.csv | A multivariate classification time-series dataset consists of 6711 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.36 showing a Unbalanced dataset. Among the 6711 samples the target ground-truth class has changed 993 times representing a percentage of 14.87%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 13 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.004. to 17.221 and kurtosis values of 0.83 to 355.42. The fractal dimension analysis yields values ranging from -0.64 to -0.11 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.94, maximum 1.00, mean 0.09, and standard deviation 0.48. The count of numerical predictors with outliers is 16 with the minimum percentage of 2.19%, maximum percentage of 48.32%, average percentage of 36.66%, and standard deviation percentage of 13.11%.
The dataset is converted into a simple classification task by extracting the previously described features. | ElasticNetClassifier | {'C': 1000.0, 'l1_ratio': 0.00055, 'penalty': 'elasticnet', 'solver': 'saga'} |
1003-classification.csv | A multivariate classification time-series dataset consists of 299 samples and 1 features with 1 numerical and 0 categorical features. Each instance has a window length of 7. The dataset has a sampling rate of 1440.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00
The target column has 3 classes with entropy value 1.33 showing a Unbalanced dataset. Among the 299 samples the target ground-truth class has changed 123 times representing a percentage of 43.62%. There are 1 features in the dataset
Among the numerical predictors, the series has 1 numerical features detected as Stationary out of the 1 numerical features using the dickey-fuller test and the rest are Unstationary. 1 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 1 seasonality components detected in the numerical predictors. The top 1 common seasonality components are represented using sinusoidal waves. of periods 6 The numerical predictors also exhibit skewness values ranging from 2.584. to 2.584 and kurtosis values of 18.95 to 18.95. The fractal dimension analysis yields values ranging from -0.78 to -0.78 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 1.00, maximum 1.00, mean 1.00, and standard deviation 0.00. The count of numerical predictors with outliers is 1 with the minimum percentage of 1.77%, maximum percentage of 1.77%, average percentage of 1.77%,
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=5), 'learning_rate': 0.01, 'n_estimators': 150} |
1031-39-2-1-1-classification.csv | A multivariate classification time-series dataset consists of 7221 samples and 16 features with 15 numerical and 1 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0
The target column has 3 classes with entropy value 1.37 showing a Unbalanced dataset. Among the 7221 samples the target ground-truth class has changed 1253 times representing a percentage of 17.43%. There are 16 features in the dataset with a ratio of numerical to categorical features of 15.0.
Among the numerical predictors, the series has 12 numerical features detected as Stationary out of the 15 numerical features using the dickey-fuller test and the rest are Unstationary. 12 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.002. to 1.393 and kurtosis values of 0.42 to 4.30. The fractal dimension analysis yields values ranging from -0.59 to -0.13 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.82, maximum 1.00, mean 0.10, and standard deviation 0.56. The count of numerical predictors with outliers is 15 with the minimum percentage of 12.73%, maximum percentage of 32.48%, average percentage of 25.75%, and standard deviation percentage of 8.03%.
Among the categorical predictors, the count of symbols ranges from 76 to 76 with a minimum entropy value 0.3588927844007235, maximum entropy 0.3588927844007235, mean 0.3588927844007235, and standard deviation 0.0,
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 10, 'n_estimators': 50} |
1031-3-1-1-5-classification.csv | A multivariate classification time-series dataset consists of 6404 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.48 showing a Unbalanced dataset. Among the 6404 samples the target ground-truth class has changed 1429 times representing a percentage of 22.43%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 13 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.012. to 10.361 and kurtosis values of 0.15 to 134.85. The fractal dimension analysis yields values ranging from -0.76 to -0.15 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.97, maximum 1.00, mean 0.10, and standard deviation 0.48. The count of numerical predictors with outliers is 16 with the minimum percentage of 0.11%, maximum percentage of 15.65%, average percentage of 5.40%, and standard deviation percentage of 4.09%.
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
1031-42-1-1-4-classification.csv | A multivariate classification time-series dataset consists of 7376 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.47 showing a Unbalanced dataset. Among the 7376 samples the target ground-truth class has changed 1878 times representing a percentage of 25.58%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 14 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.035. to 8.219 and kurtosis values of 0.13 to 105.21. The fractal dimension analysis yields values ranging from -0.69 to -0.13 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.96, maximum 1.00, mean 0.12, and standard deviation 0.49. The count of numerical predictors with outliers is 16 with the minimum percentage of 1.51%, maximum percentage of 12.09%, average percentage of 6.78%, and standard deviation percentage of 2.83%.
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
1031-13-2-1-3-classification.csv | A multivariate classification time-series dataset consists of 6381 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.36 showing a Unbalanced dataset. Among the 6381 samples the target ground-truth class has changed 787 times representing a percentage of 12.40%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 16 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.094. to 20.779 and kurtosis values of 1.00 to 630.09. The fractal dimension analysis yields values ranging from -0.68 to -0.11 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.96, maximum 1.00, mean 0.10, and standard deviation 0.48. The count of numerical predictors with outliers is 16 with the minimum percentage of 2.30%, maximum percentage of 43.69%, average percentage of 38.81%, and standard deviation percentage of 10.32%.
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
1016-5-2-5-classification.csv | A multivariate classification time-series dataset consists of 1593 samples and 8 features with 8 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 2 classes with entropy value 1.00 showing a Balanced dataset. Among the 1593 samples the target ground-truth class has changed 92 times representing a percentage of 5.90%. There are 8 features in the dataset
Among the numerical predictors, the series has 8 numerical features detected as Stationary out of the 8 numerical features using the dickey-fuller test and the rest are Unstationary. 7 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 2 seasonality components detected in the numerical predictors. The top 2 common seasonality components are represented using sinusoidal waves. of periods 19,24 The numerical predictors also exhibit skewness values ranging from 0.007. to 5.362 and kurtosis values of 0.18 to 43.28. The fractal dimension analysis yields values ranging from -0.66 to -0.09 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.51, maximum 1.00, mean 0.29, and standard deviation 0.44. The count of numerical predictors with outliers is 4 with the minimum percentage of 0.00%, maximum percentage of 18.73%, average percentage of 3.51%, and standard deviation percentage of 6.58%.
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 20, 'n_estimators': 100} |
1031-25-2-1-5-classification.csv | A multivariate classification time-series dataset consists of 3252 samples and 16 features with 15 numerical and 1 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0
The target column has 3 classes with entropy value 1.43 showing a Unbalanced dataset. Among the 3252 samples the target ground-truth class has changed 675 times representing a percentage of 20.98%. There are 16 features in the dataset with a ratio of numerical to categorical features of 15.0.
Among the numerical predictors, the series has 14 numerical features detected as Stationary out of the 15 numerical features using the dickey-fuller test and the rest are Unstationary. 12 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.037. to 1.651 and kurtosis values of 0.07 to 7.23. The fractal dimension analysis yields values ranging from -0.73 to -0.13 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.93, maximum 1.00, mean 0.15, and standard deviation 0.48. The count of numerical predictors with outliers is 15 with the minimum percentage of 0.31%, maximum percentage of 15.72%, average percentage of 8.31%, and standard deviation percentage of 6.35%.
Among the categorical predictors, the count of symbols ranges from 73 to 73 with a minimum entropy value 0.4549698318220819, maximum entropy 0.4549698318220819, mean 0.4549698318220819, and standard deviation 0.0,
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 1.0, 'n_estimators': 50} |
1028-5-classification.csv | A multivariate classification time-series dataset consists of 6231 samples and 8 features with 8 numerical and 0 categorical features. Each instance has a window length of 7. The dataset has a sampling rate of 1440.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.35 showing a Unbalanced dataset. Among the 6231 samples the target ground-truth class has changed 10 times representing a percentage of 0.16%. There are 8 features in the dataset
Among the numerical predictors, the series has 2 numerical features detected as Stationary out of the 8 numerical features using the dickey-fuller test and the rest are Unstationary. 8 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 2.033. to 18.232 and kurtosis values of 2.69 to 524.05. The fractal dimension analysis yields values ranging from -0.60 to -0.31 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 0.15, maximum 1.00, mean 0.74, and standard deviation 0.34. The count of numerical predictors with outliers is 8 with the minimum percentage of 8.19%, maximum percentage of 18.75%, average percentage of 16.02%, and standard deviation percentage of 3.23%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=5), 'learning_rate': 0.01, 'n_estimators': 100} |
3001-2.csv | A multivariate classification time-series dataset consists of 672 samples and 2 features with 2 numerical and 0 categorical features. Each instance has a window length of 4. The dataset has a sampling rate of 360.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.50 showing a Unbalanced dataset. Among the 672 samples the target ground-truth class has changed 655 times representing a percentage of 99.54%. There are 2 features in the dataset
Among the numerical predictors, the series has 0 numerical features detected as Stationary out of the 2 numerical features using the dickey-fuller test and the rest are Unstationary. 0 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.013. to 0.184 and kurtosis values of 0.16 to 0.47. The fractal dimension analysis yields values ranging from -1.50 to -1.21 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.01, maximum 1.00, mean 0.49, and standard deviation 0.51. The count of numerical predictors with outliers is 1 with the minimum percentage of 0.00%, maximum percentage of 0.76%, average percentage of 0.38%, and standard deviation percentage of 0.54%.
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 10, 'n_estimators': 100} |
1031-36-1-1-2-classification.csv | A multivariate classification time-series dataset consists of 7680 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.35 showing a Unbalanced dataset. Among the 7680 samples the target ground-truth class has changed 354 times representing a percentage of 4.63%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 16 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.517. to 17.442 and kurtosis values of 2.59 to 306.16. The fractal dimension analysis yields values ranging from -0.42 to -0.09 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 0.07, maximum 1.00, mean 0.88, and standard deviation 0.28. The count of numerical predictors with outliers is 16 with the minimum percentage of 20.22%, maximum percentage of 20.22%, average percentage of 20.22%, and standard deviation percentage of 0.00%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=1), 'learning_rate': 0.1, 'n_estimators': 50} |
3001-57.csv | A multivariate classification time-series dataset consists of 480 samples and 1 features with 1 numerical and 0 categorical features. Each instance has a window length of 4. The dataset has a sampling rate of 360.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00
The target column has 3 classes with entropy value 1.34 showing a Unbalanced dataset. Among the 480 samples the target ground-truth class has changed 353 times representing a percentage of 75.59%. There are 1 features in the dataset
Among the numerical predictors, the series has 1 numerical features detected as Stationary out of the 1 numerical features using the dickey-fuller test and the rest are Unstationary. 1 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 1.257. to 1.257 and kurtosis values of 0.78 to 0.78. The fractal dimension analysis yields values ranging from -1.18 to -1.18 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 1.00, maximum 1.00, mean 1.00, and standard deviation 0.00. The count of numerical predictors with outliers is 1 with the minimum percentage of 4.07%, maximum percentage of 4.07%, average percentage of 4.07%,
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=5), 'learning_rate': 2.0, 'n_estimators': 100} |
3001-54.csv | A multivariate classification time-series dataset consists of 720 samples and 1 features with 1 numerical and 0 categorical features. Each instance has a window length of 3. The dataset has a sampling rate of 480.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00
The target column has 4 classes with entropy value 1.25 showing a Unbalanced dataset. Among the 720 samples the target ground-truth class has changed 473 times representing a percentage of 66.81%. There are 1 features in the dataset
Among the numerical predictors, the series has 0 numerical features detected as Stationary out of the 1 numerical features using the dickey-fuller test and the rest are Unstationary. 1 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.445. to 0.445 and kurtosis values of 0.60 to 0.60. The fractal dimension analysis yields values ranging from -1.22 to -1.22 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 1.00, maximum 1.00, mean 1.00, and standard deviation 0.00. The count of numerical predictors with outliers is 1 with the minimum percentage of 0.14%, maximum percentage of 0.14%, average percentage of 0.14%,
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=1), 'learning_rate': 0.01, 'n_estimators': 50} |
1031-6-2-1-2-classification.csv | A multivariate classification time-series dataset consists of 5177 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.34 showing a Unbalanced dataset. Among the 5177 samples the target ground-truth class has changed 250 times representing a percentage of 4.86%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 16 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.280. to 20.339 and kurtosis values of 2.75 to 415.90. The fractal dimension analysis yields values ranging from -0.37 to -0.04 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 0.06, maximum 1.00, mean 0.87, and standard deviation 0.28. The count of numerical predictors with outliers is 16 with the minimum percentage of 19.39%, maximum percentage of 19.39%, average percentage of 19.39%, and standard deviation percentage of 0.00%.
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 10, 'n_estimators': 50} |
1030-272-classification.csv | A multivariate classification time-series dataset consists of 4140 samples and 5 features with 5 numerical and 0 categorical features. Each instance has a window length of 7. The dataset has a sampling rate of 1440.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.53 showing a Unbalanced dataset. Among the 4140 samples the target ground-truth class has changed 87 times representing a percentage of 2.11%. There are 5 features in the dataset
Among the numerical predictors, the series has 1 numerical features detected as Stationary out of the 5 numerical features using the dickey-fuller test and the rest are Unstationary. 5 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.142. to 2.227 and kurtosis values of 0.54 to 8.56. The fractal dimension analysis yields values ranging from -0.62 to -0.31 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.13, maximum 1.00, mean 0.63, and standard deviation 0.49. The count of numerical predictors with outliers is 4 with the minimum percentage of 0.00%, maximum percentage of 4.92%, average percentage of 1.03%, and standard deviation percentage of 2.18%.
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
1031-21-1-1-5-classification.csv | A multivariate classification time-series dataset consists of 7508 samples and 13 features with 13 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.50 showing a Unbalanced dataset. Among the 7508 samples the target ground-truth class has changed 1586 times representing a percentage of 21.22%. There are 13 features in the dataset
Among the numerical predictors, the series has 13 numerical features detected as Stationary out of the 13 numerical features using the dickey-fuller test and the rest are Unstationary. 9 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.006. to 14.056 and kurtosis values of 0.18 to 259.00. The fractal dimension analysis yields values ranging from -0.70 to -0.14 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.96, maximum 1.00, mean 0.18, and standard deviation 0.53. The count of numerical predictors with outliers is 13 with the minimum percentage of 0.01%, maximum percentage of 10.69%, average percentage of 4.12%, and standard deviation percentage of 4.06%.
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
1031-28-1-1-5-classification.csv | A multivariate classification time-series dataset consists of 7612 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.44 showing a Unbalanced dataset. Among the 7612 samples the target ground-truth class has changed 1558 times representing a percentage of 20.56%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 14 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.044. to 11.256 and kurtosis values of 0.18 to 156.36. The fractal dimension analysis yields values ranging from -0.62 to -0.12 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.90, maximum 1.00, mean 0.15, and standard deviation 0.46. The count of numerical predictors with outliers is 16 with the minimum percentage of 0.80%, maximum percentage of 16.39%, average percentage of 7.98%, and standard deviation percentage of 4.78%.
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
1016-4-1-4-classification.csv | A multivariate classification time-series dataset consists of 7109 samples and 12 features with 4 numerical and 8 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0 and standard deviation 0.0.
The target column has 2 classes with entropy value 1.00 showing a Balanced dataset. Among the 7109 samples the target ground-truth class has changed 324 times representing a percentage of 4.58%. There are 12 features in the dataset with a ratio of numerical to categorical features of 0.5.
Among the numerical predictors, the series has 4 numerical features detected as Stationary out of the 4 numerical features using the dickey-fuller test and the rest are Unstationary. 4 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 3 seasonality components detected in the numerical predictors. The top 3 common seasonality components are represented using sinusoidal waves. of periods 89,161,182 The numerical predictors also exhibit skewness values ranging from 0.060. to 0.466 and kurtosis values of 0.13 to 0.42. The fractal dimension analysis yields values ranging from -0.52 to -0.09 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.14, maximum 1.00, mean 0.56, and standard deviation 0.54. The count of numerical predictors with outliers is 4 with the minimum percentage of 0.06%, maximum percentage of 0.89%, average percentage of 0.39%, and standard deviation percentage of 0.36%.
Among the categorical predictors, the count of symbols ranges from 9 to 69 with a minimum entropy value 1.2981849496968765, maximum entropy 5.192394790748429, mean 3.6678600330556614, and standard deviation 1.5032559852031107,
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 1.0, 'n_estimators': 50} |
1030-312-classification.csv | A multivariate classification time-series dataset consists of 4140 samples and 5 features with 5 numerical and 0 categorical features. Each instance has a window length of 7. The dataset has a sampling rate of 1440.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.55 showing a Unbalanced dataset. Among the 4140 samples the target ground-truth class has changed 26 times representing a percentage of 0.63%. There are 5 features in the dataset
Among the numerical predictors, the series has 1 numerical features detected as Stationary out of the 5 numerical features using the dickey-fuller test and the rest are Unstationary. 5 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.230. to 2.986 and kurtosis values of 0.93 to 14.98. The fractal dimension analysis yields values ranging from -0.57 to -0.33 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.35, maximum 1.00, mean 0.57, and standard deviation 0.63. The count of numerical predictors with outliers is 1 with the minimum percentage of 0.00%, maximum percentage of 5.17%, average percentage of 1.03%, and standard deviation percentage of 2.31%.
The dataset is converted into a simple classification task by extracting the previously described features. | ElasticNetClassifier | {'C': 1000.0, 'l1_ratio': 0.0001, 'penalty': 'elasticnet', 'solver': 'saga'} |
1031-22-1-1-6-classification.csv | A multivariate classification time-series dataset consists of 7607 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.37 showing a Unbalanced dataset. Among the 7607 samples the target ground-truth class has changed 518 times representing a percentage of 6.84%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 16 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.711. to 9.533 and kurtosis values of 1.24 to 94.19. The fractal dimension analysis yields values ranging from -0.17 to -0.08 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 0.22, maximum 1.00, mean 0.59, and standard deviation 0.37. The count of numerical predictors with outliers is 16 with the minimum percentage of 0.91%, maximum percentage of 27.64%, average percentage of 12.61%, and standard deviation percentage of 13.69%.
The dataset is converted into a simple classification task by extracting the previously described features. | ElasticNetClassifier | {'C': 181.8181818181818, 'l1_ratio': 0.00055, 'penalty': 'elasticnet', 'solver': 'saga'} |
1031-4-3-1-5-classification.csv | A multivariate classification time-series dataset consists of 7651 samples and 16 features with 15 numerical and 1 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0
The target column has 3 classes with entropy value 1.41 showing a Unbalanced dataset. Among the 7651 samples the target ground-truth class has changed 1415 times representing a percentage of 18.58%. There are 16 features in the dataset with a ratio of numerical to categorical features of 15.0.
Among the numerical predictors, the series has 15 numerical features detected as Stationary out of the 15 numerical features using the dickey-fuller test and the rest are Unstationary. 12 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.000. to 1.520 and kurtosis values of 0.10 to 3.16. The fractal dimension analysis yields values ranging from -0.63 to -0.13 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.85, maximum 1.00, mean 0.12, and standard deviation 0.55. The count of numerical predictors with outliers is 15 with the minimum percentage of 0.45%, maximum percentage of 15.07%, average percentage of 7.50%, and standard deviation percentage of 4.01%.
Among the categorical predictors, the count of symbols ranges from 113 to 113 with a minimum entropy value 0.49680385088795465, maximum entropy 0.49680385088795465, mean 0.49680385088795465, and standard deviation 0.0,
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
1031-39-1-1-4-classification.csv | A multivariate classification time-series dataset consists of 7594 samples and 15 features with 15 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.43 showing a Unbalanced dataset. Among the 7594 samples the target ground-truth class has changed 1454 times representing a percentage of 19.23%. There are 15 features in the dataset
Among the numerical predictors, the series has 12 numerical features detected as Stationary out of the 15 numerical features using the dickey-fuller test and the rest are Unstationary. 14 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.045. to 1.829 and kurtosis values of 0.08 to 3.31. The fractal dimension analysis yields values ranging from -0.58 to -0.13 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.88, maximum 1.00, mean 0.12, and standard deviation 0.55. The count of numerical predictors with outliers is 15 with the minimum percentage of 0.97%, maximum percentage of 13.25%, average percentage of 7.64%, and standard deviation percentage of 3.99%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 1.0, 'n_estimators': 50} |
1030-308-classification.csv | A multivariate classification time-series dataset consists of 4140 samples and 5 features with 5 numerical and 0 categorical features. Each instance has a window length of 7. The dataset has a sampling rate of 1440.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 4 classes with entropy value 1.98 showing a Unbalanced dataset. Among the 4140 samples the target ground-truth class has changed 21 times representing a percentage of 0.51%. There are 5 features in the dataset
Among the numerical predictors, the series has 1 numerical features detected as Stationary out of the 5 numerical features using the dickey-fuller test and the rest are Unstationary. 5 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.797. to 3.685 and kurtosis values of 0.43 to 32.64. The fractal dimension analysis yields values ranging from -0.65 to -0.31 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 0.10, maximum 1.00, mean 0.71, and standard deviation 0.42. The count of numerical predictors with outliers is 5 with the minimum percentage of 0.41%, maximum percentage of 4.88%, average percentage of 1.77%, and standard deviation percentage of 1.77%.
The dataset is converted into a simple classification task by extracting the previously described features. | ElasticNetClassifier | {'C': 1000.0, 'l1_ratio': 0.0001, 'penalty': 'elasticnet', 'solver': 'saga'} |
1031-41-1-1-5-classification.csv | A multivariate classification time-series dataset consists of 7612 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.43 showing a Unbalanced dataset. Among the 7612 samples the target ground-truth class has changed 1507 times representing a percentage of 19.89%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 14 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.015. to 15.574 and kurtosis values of 0.06 to 331.52. The fractal dimension analysis yields values ranging from -0.70 to -0.12 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.98, maximum 1.00, mean 0.12, and standard deviation 0.49. The count of numerical predictors with outliers is 16 with the minimum percentage of 2.90%, maximum percentage of 16.53%, average percentage of 8.02%, and standard deviation percentage of 4.42%.
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
1031-8-2-1-2-classification.csv | A multivariate classification time-series dataset consists of 7121 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.37 showing a Unbalanced dataset. Among the 7121 samples the target ground-truth class has changed 1162 times representing a percentage of 16.40%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 15 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.017. to 22.516 and kurtosis values of 0.77 to 605.91. The fractal dimension analysis yields values ranging from -0.66 to -0.13 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.95, maximum 1.00, mean 0.11, and standard deviation 0.54. The count of numerical predictors with outliers is 16 with the minimum percentage of 1.88%, maximum percentage of 47.45%, average percentage of 40.79%, and standard deviation percentage of 13.52%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 1.0, 'n_estimators': 50} |
1020-28-3-classification.csv | A multivariate classification time-series dataset consists of 7012 samples and 11 features with 9 numerical and 2 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0 and standard deviation 0.0.
The target column has 2 classes with entropy value 0.97 showing a Unbalanced dataset. Among the 7012 samples the target ground-truth class has changed 636 times representing a percentage of 9.11%. There are 11 features in the dataset with a ratio of numerical to categorical features of 4.5.
Among the numerical predictors, the series has 8 numerical features detected as Stationary out of the 9 numerical features using the dickey-fuller test and the rest are Unstationary. 9 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 44 seasonality components detected in the numerical predictors. The top 3 common seasonality components are represented using sinusoidal waves. of periods 1402,1753,2337 The numerical predictors also exhibit skewness values ranging from 0.310. to 3.687 and kurtosis values of 0.26 to 20.02. The fractal dimension analysis yields values ranging from -0.66 to -0.14 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.81, maximum 1.00, mean 0.16, and standard deviation 0.48. The count of numerical predictors with outliers is 7 with the minimum percentage of 0.00%, maximum percentage of 7.91%, average percentage of 3.61%, and standard deviation percentage of 2.99%.
Among the categorical predictors, the count of symbols ranges from 17 to 66 with a minimum entropy value 0.6193731453399517, maximum entropy 3.9083400080671242, mean 2.263856576703538, and standard deviation 1.6444834313635863,
The dataset is converted into a simple classification task by extracting the previously described features. | LassoClassifier | {'C': 640.428052016275, 'penalty': 'l1', 'solver': 'saga'} |
1031-44-2-1-5-classification.csv | A multivariate classification time-series dataset consists of 7996 samples and 16 features with 13 numerical and 3 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 1791 with mean 137.77 and standard deviation 496.73. Similarly, the missing values percentages for categorical features range from 0 to 1791 with mean 1194.0 and standard deviation 1034.0343321186197.
The target column has 3 classes with entropy value 1.54 showing a Unbalanced dataset. Among the 7996 samples the target ground-truth class has changed 1122 times representing a percentage of 18.18%. There are 16 features in the dataset with a ratio of numerical to categorical features of 4.333333333333333.
Among the numerical predictors, the series has 13 numerical features detected as Stationary out of the 13 numerical features using the dickey-fuller test and the rest are Unstationary. 13 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.079. to 1.168 and kurtosis values of 0.04 to 2.85. The fractal dimension analysis yields values ranging from -0.69 to -0.13 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.96, maximum 1.00, mean 0.11, and standard deviation 0.51. The count of numerical predictors with outliers is 12 with the minimum percentage of 0.00%, maximum percentage of 3.50%, average percentage of 0.95%, and standard deviation percentage of 1.11%.
Among the categorical predictors, the count of symbols ranges from 27 to 77 with a minimum entropy value 0.1319481767880569, maximum entropy 5.061439052615287, mean 3.249863218931521, and standard deviation 2.2143263338820542,
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 1.0, 'n_estimators': 50} |
1016-16-1-3-classification.csv | A multivariate classification time-series dataset consists of 7564 samples and 6 features with 5 numerical and 1 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 276 to 276 with mean 276.0
The target column has 2 classes with entropy value 0.99 showing a Balanced dataset. Among the 7564 samples the target ground-truth class has changed 328 times representing a percentage of 4.52%. There are 6 features in the dataset with a ratio of numerical to categorical features of 5.0.
Among the numerical predictors, the series has 5 numerical features detected as Stationary out of the 5 numerical features using the dickey-fuller test and the rest are Unstationary. 4 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 10 seasonality components detected in the numerical predictors. The top 3 common seasonality components are represented using sinusoidal waves. of periods 328,455,911 The numerical predictors also exhibit skewness values ranging from 0.019. to 0.741 and kurtosis values of 0.05 to 1.04. The fractal dimension analysis yields values ranging from -0.53 to -0.12 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.38, maximum 1.00, mean 0.32, and standard deviation 0.46. The count of numerical predictors with outliers is 4 with the minimum percentage of 0.00%, maximum percentage of 1.08%, average percentage of 0.41%, and standard deviation percentage of 0.52%.
Among the categorical predictors, the count of symbols ranges from 44 to 44 with a minimum entropy value 0.9994460798643348, maximum entropy 0.9994460798643348, mean 0.9994460798643348, and standard deviation 0.0,
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 20, 'n_estimators': 250} |
1031-20-2-1-4-classification.csv | A multivariate classification time-series dataset consists of 6981 samples and 16 features with 15 numerical and 1 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0
The target column has 3 classes with entropy value 1.35 showing a Unbalanced dataset. Among the 6981 samples the target ground-truth class has changed 388 times representing a percentage of 5.59%. There are 16 features in the dataset with a ratio of numerical to categorical features of 15.0.
Among the numerical predictors, the series has 15 numerical features detected as Stationary out of the 15 numerical features using the dickey-fuller test and the rest are Unstationary. 15 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.222. to 1.890 and kurtosis values of 3.71 to 14.31. The fractal dimension analysis yields values ranging from -0.49 to -0.12 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.98, maximum 1.00, mean 0.13, and standard deviation 0.57. The count of numerical predictors with outliers is 15 with the minimum percentage of 25.84%, maximum percentage of 25.84%, average percentage of 25.84%, and standard deviation percentage of 0.00%.
Among the categorical predictors, the count of symbols ranges from 23 to 23 with a minimum entropy value 0.060044854129450494, maximum entropy 0.060044854129450494, mean 0.060044854129450494, and standard deviation 0.0,
The dataset is converted into a simple classification task by extracting the previously described features. | ElasticNetClassifier | {'C': 1000.0, 'l1_ratio': 0.0001, 'penalty': 'elasticnet', 'solver': 'saga'} |
1016-22-1-5-classification.csv | A multivariate classification time-series dataset consists of 7109 samples and 8 features with 8 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.46 showing a Unbalanced dataset. Among the 7109 samples the target ground-truth class has changed 796 times representing a percentage of 11.25%. There are 8 features in the dataset
Among the numerical predictors, the series has 8 numerical features detected as Stationary out of the 8 numerical features using the dickey-fuller test and the rest are Unstationary. 8 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 37 seasonality components detected in the numerical predictors. The top 3 common seasonality components are represented using sinusoidal waves. of periods 546,592,1184 The numerical predictors also exhibit skewness values ranging from 0.206. to 10.037 and kurtosis values of 0.01 to 174.02. The fractal dimension analysis yields values ranging from -0.59 to -0.08 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.39, maximum 1.00, mean 0.18, and standard deviation 0.45. The count of numerical predictors with outliers is 7 with the minimum percentage of 0.00%, maximum percentage of 8.83%, average percentage of 2.67%, and standard deviation percentage of 2.78%.
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 10, 'n_estimators': 30} |
3001-24.csv | A multivariate classification time-series dataset consists of 480 samples and 1 features with 1 numerical and 0 categorical features. Each instance has a window length of 4. The dataset has a sampling rate of 360.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00
The target column has 3 classes with entropy value 1.33 showing a Unbalanced dataset. Among the 480 samples the target ground-truth class has changed 354 times representing a percentage of 75.64%. There are 1 features in the dataset
Among the numerical predictors, the series has 0 numerical features detected as Stationary out of the 1 numerical features using the dickey-fuller test and the rest are Unstationary. 1 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.392. to 0.392 and kurtosis values of 1.00 to 1.00. The fractal dimension analysis yields values ranging from -1.80 to -1.80 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 1.00, maximum 1.00, mean 1.00, and standard deviation 0.00. The count of numerical predictors with outliers is 0 with the minimum percentage of 0.00%, maximum percentage of 0.00%, average percentage of 0.00%,
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=1), 'learning_rate': 0.01, 'n_estimators': 50} |
1031-54-2-1-1-classification.csv | A multivariate classification time-series dataset consists of 6916 samples and 16 features with 15 numerical and 1 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0
The target column has 3 classes with entropy value 1.37 showing a Unbalanced dataset. Among the 6916 samples the target ground-truth class has changed 1171 times representing a percentage of 17.02%. There are 16 features in the dataset with a ratio of numerical to categorical features of 15.0.
Among the numerical predictors, the series has 15 numerical features detected as Stationary out of the 15 numerical features using the dickey-fuller test and the rest are Unstationary. 12 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.018. to 2.164 and kurtosis values of 0.33 to 9.74. The fractal dimension analysis yields values ranging from -0.55 to -0.12 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.94, maximum 1.00, mean 0.12, and standard deviation 0.54. The count of numerical predictors with outliers is 15 with the minimum percentage of 14.46%, maximum percentage of 35.43%, average percentage of 29.05%, and standard deviation percentage of 5.13%.
Among the categorical predictors, the count of symbols ranges from 54 to 54 with a minimum entropy value 0.21562894264852864, maximum entropy 0.21562894264852864, mean 0.21562894264852864, and standard deviation 0.0,
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 10, 'n_estimators': 30} |
1031-27-2-1-5-classification.csv | A multivariate classification time-series dataset consists of 7881 samples and 12 features with 12 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.36 showing a Unbalanced dataset. Among the 7881 samples the target ground-truth class has changed 680 times representing a percentage of 8.67%. There are 12 features in the dataset
Among the numerical predictors, the series has 12 numerical features detected as Stationary out of the 12 numerical features using the dickey-fuller test and the rest are Unstationary. 10 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.051. to 2.390 and kurtosis values of 1.86 to 11.88. The fractal dimension analysis yields values ranging from -0.38 to -0.13 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.95, maximum 1.00, mean 0.17, and standard deviation 0.55. The count of numerical predictors with outliers is 12 with the minimum percentage of 39.67%, maximum percentage of 39.67%, average percentage of 39.67%, and standard deviation percentage of 0.00%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 1.0, 'n_estimators': 50} |
1031-49-2-1-1-classification.csv | A multivariate classification time-series dataset consists of 6836 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.46 showing a Unbalanced dataset. Among the 6836 samples the target ground-truth class has changed 1522 times representing a percentage of 22.38%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 15 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.006. to 19.572 and kurtosis values of 0.04 to 502.83. The fractal dimension analysis yields values ranging from -0.71 to -0.13 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.88, maximum 1.00, mean 0.10, and standard deviation 0.50. The count of numerical predictors with outliers is 16 with the minimum percentage of 1.48%, maximum percentage of 13.91%, average percentage of 7.38%, and standard deviation percentage of 4.19%.
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
1031-9-2-1-1-classification.csv | A multivariate classification time-series dataset consists of 7900 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.50 showing a Unbalanced dataset. Among the 7900 samples the target ground-truth class has changed 1674 times representing a percentage of 21.28%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 16 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.076. to 20.582 and kurtosis values of 0.20 to 576.20. The fractal dimension analysis yields values ranging from -0.76 to -0.11 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.97, maximum 1.00, mean 0.15, and standard deviation 0.50. The count of numerical predictors with outliers is 16 with the minimum percentage of 0.01%, maximum percentage of 13.30%, average percentage of 4.00%, and standard deviation percentage of 4.08%.
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
1031-31-2-1-4-classification.csv | A multivariate classification time-series dataset consists of 7778 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.36 showing a Unbalanced dataset. Among the 7778 samples the target ground-truth class has changed 1211 times representing a percentage of 15.64%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 15 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.056. to 12.922 and kurtosis values of 0.13 to 266.82. The fractal dimension analysis yields values ranging from -0.67 to -0.16 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.94, maximum 1.00, mean 0.10, and standard deviation 0.50. The count of numerical predictors with outliers is 15 with the minimum percentage of 0.00%, maximum percentage of 13.33%, average percentage of 7.97%, and standard deviation percentage of 3.62%.
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
1031-6-3-1-2-classification.csv | A multivariate classification time-series dataset consists of 6646 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.41 showing a Unbalanced dataset. Among the 6646 samples the target ground-truth class has changed 1227 times representing a percentage of 18.56%. There are 16 features in the dataset
Among the numerical predictors, the series has 13 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 15 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.099. to 17.461 and kurtosis values of 0.01 to 419.39. The fractal dimension analysis yields values ranging from -0.63 to -0.11 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.83, maximum 1.00, mean 0.09, and standard deviation 0.53. The count of numerical predictors with outliers is 15 with the minimum percentage of 0.00%, maximum percentage of 11.87%, average percentage of 5.25%, and standard deviation percentage of 4.32%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 1.0, 'n_estimators': 50} |
1031-81-1-4-classification.csv | A multivariate classification time-series dataset consists of 7800 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.35 showing a Unbalanced dataset. Among the 7800 samples the target ground-truth class has changed 443 times representing a percentage of 5.70%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 16 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.108. to 21.443 and kurtosis values of 1.94 to 568.97. The fractal dimension analysis yields values ranging from -0.70 to -0.17 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.95, maximum 1.00, mean 0.10, and standard deviation 0.60. The count of numerical predictors with outliers is 16 with the minimum percentage of 1.62%, maximum percentage of 33.20%, average percentage of 31.22%, and standard deviation percentage of 7.89%.
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 10, 'n_estimators': 400} |
1031-60-2-1-1-classification.csv | A multivariate classification time-series dataset consists of 6615 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.41 showing a Unbalanced dataset. Among the 6615 samples the target ground-truth class has changed 1070 times representing a percentage of 16.26%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 16 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.161. to 11.957 and kurtosis values of 0.19 to 187.80. The fractal dimension analysis yields values ranging from -0.74 to -0.14 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.97, maximum 1.00, mean 0.10, and standard deviation 0.47. The count of numerical predictors with outliers is 16 with the minimum percentage of 4.19%, maximum percentage of 25.09%, average percentage of 13.85%, and standard deviation percentage of 5.22%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 1.0, 'n_estimators': 50} |
1031-18-1-1-1-classification.csv | A multivariate classification time-series dataset consists of 7362 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.41 showing a Unbalanced dataset. Among the 7362 samples the target ground-truth class has changed 1287 times representing a percentage of 17.56%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 15 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.023. to 12.664 and kurtosis values of 0.36 to 220.72. The fractal dimension analysis yields values ranging from -0.65 to -0.14 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.88, maximum 1.00, mean 0.15, and standard deviation 0.43. The count of numerical predictors with outliers is 15 with the minimum percentage of 0.00%, maximum percentage of 9.59%, average percentage of 5.33%, and standard deviation percentage of 2.04%.
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 10, 'n_estimators': 50} |
1031-15-1-1-4-classification.csv | A multivariate classification time-series dataset consists of 7606 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.47 showing a Unbalanced dataset. Among the 7606 samples the target ground-truth class has changed 1618 times representing a percentage of 21.37%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 15 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.011. to 15.884 and kurtosis values of 0.00 to 321.97. The fractal dimension analysis yields values ranging from -0.66 to -0.11 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.95, maximum 1.00, mean 0.10, and standard deviation 0.48. The count of numerical predictors with outliers is 15 with the minimum percentage of 0.00%, maximum percentage of 14.41%, average percentage of 5.89%, and standard deviation percentage of 4.64%.
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
1031-16-1-1-2-classification.csv | A multivariate classification time-series dataset consists of 7459 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.41 showing a Unbalanced dataset. Among the 7459 samples the target ground-truth class has changed 1425 times representing a percentage of 19.19%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 15 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.062. to 14.159 and kurtosis values of 0.00 to 259.39. The fractal dimension analysis yields values ranging from -0.65 to -0.11 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.87, maximum 1.00, mean 0.09, and standard deviation 0.48. The count of numerical predictors with outliers is 16 with the minimum percentage of 3.33%, maximum percentage of 12.30%, average percentage of 7.82%, and standard deviation percentage of 3.28%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 1.0, 'n_estimators': 50} |
1031-7-3-1-1-classification.csv | A multivariate classification time-series dataset consists of 6728 samples and 16 features with 15 numerical and 1 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0
The target column has 3 classes with entropy value 1.45 showing a Unbalanced dataset. Among the 6728 samples the target ground-truth class has changed 1249 times representing a percentage of 18.66%. There are 16 features in the dataset with a ratio of numerical to categorical features of 15.0.
Among the numerical predictors, the series has 15 numerical features detected as Stationary out of the 15 numerical features using the dickey-fuller test and the rest are Unstationary. 10 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.004. to 1.350 and kurtosis values of 0.05 to 4.96. The fractal dimension analysis yields values ranging from -0.60 to -0.11 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.89, maximum 1.00, mean 0.11, and standard deviation 0.56. The count of numerical predictors with outliers is 15 with the minimum percentage of 0.18%, maximum percentage of 15.06%, average percentage of 7.16%, and standard deviation percentage of 5.09%.
Among the categorical predictors, the count of symbols ranges from 108 to 108 with a minimum entropy value 0.5189352208299154, maximum entropy 0.5189352208299154, mean 0.5189352208299154, and standard deviation 0.0,
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 5, 'n_estimators': 50} |
3001-18.csv | A multivariate classification time-series dataset consists of 504 samples and 3 features with 3 numerical and 0 categorical features. Each instance has a window length of 3. The dataset has a sampling rate of 480.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.15 showing a Unbalanced dataset. Among the 504 samples the target ground-truth class has changed 27 times representing a percentage of 5.50%. There are 3 features in the dataset
Among the numerical predictors, the series has 0 numerical features detected as Stationary out of the 3 numerical features using the dickey-fuller test and the rest are Unstationary. 3 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.872. to 1.745 and kurtosis values of 0.42 to 2.42. The fractal dimension analysis yields values ranging from -1.33 to -1.05 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.82, maximum 1.00, mean 0.15, and standard deviation 0.78. The count of numerical predictors with outliers is 1 with the minimum percentage of 0.00%, maximum percentage of 9.78%, average percentage of 3.26%, and standard deviation percentage of 5.64%.
The dataset is converted into a simple classification task by extracting the previously described features. | GaussianProcessClassifier | {'kernel': DotProduct(sigma_0=10)} |
1016-13-5-1-classification.csv | A multivariate classification time-series dataset consists of 6595 samples and 8 features with 4 numerical and 4 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0 and standard deviation 0.0.
The target column has 2 classes with entropy value 0.97 showing a Unbalanced dataset. Among the 6595 samples the target ground-truth class has changed 283 times representing a percentage of 4.31%. There are 8 features in the dataset with a ratio of numerical to categorical features of 1.0.
Among the numerical predictors, the series has 4 numerical features detected as Stationary out of the 4 numerical features using the dickey-fuller test and the rest are Unstationary. 2 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 2 seasonality components detected in the numerical predictors. The top 2 common seasonality components are represented using sinusoidal waves. of periods 117,178 The numerical predictors also exhibit skewness values ranging from 0.024. to 0.712 and kurtosis values of 0.01 to 0.46. The fractal dimension analysis yields values ranging from -0.44 to -0.10 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.29, maximum 1.00, mean 0.51, and standard deviation 0.61. The count of numerical predictors with outliers is 4 with the minimum percentage of 0.03%, maximum percentage of 2.33%, average percentage of 1.40%, and standard deviation percentage of 0.97%.
Among the categorical predictors, the count of symbols ranges from 42 to 59 with a minimum entropy value 1.0706584762294218, maximum entropy 5.37786930392954, mean 4.0100206125484785, and standard deviation 1.7188643565958313,
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 40, 'n_estimators': 100} |
3001-31.csv | A multivariate classification time-series dataset consists of 192 samples and 1 features with 1 numerical and 0 categorical features. Each instance has a window length of 7. The dataset has a sampling rate of 1440.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00
The target column has 2 classes with entropy value 1.00 showing a Balanced dataset. Among the 192 samples the target ground-truth class has changed 2 times representing a percentage of 1.10%. There are 1 features in the dataset
Among the numerical predictors, the series has 0 numerical features detected as Stationary out of the 1 numerical features using the dickey-fuller test and the rest are Unstationary. 1 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.024. to 0.024 and kurtosis values of 1.95 to 1.95. The fractal dimension analysis yields values ranging from -0.81 to -0.81 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of 1.00, maximum 1.00, mean 1.00, and standard deviation 0.00. The count of numerical predictors with outliers is 0 with the minimum percentage of 0.00%, maximum percentage of 0.00%, average percentage of 0.00%,
The dataset is converted into a simple classification task by extracting the previously described features. | LightgbmClassifier | {'boosting_type': 'gbdt', 'colsample_bytree': 0.8, 'learning_rate': 0.01, 'max_depth': 3, 'n_estimators': 30, 'num_leaves': 50, 'reg_lambda ': 0.01} |
1031-37-1-1-5-classification.csv | A multivariate classification time-series dataset consists of 7556 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.43 showing a Unbalanced dataset. Among the 7556 samples the target ground-truth class has changed 622 times representing a percentage of 8.27%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 16 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.509. to 11.651 and kurtosis values of 0.31 to 135.79. The fractal dimension analysis yields values ranging from -0.63 to -0.07 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.02, maximum 1.00, mean 0.81, and standard deviation 0.32. The count of numerical predictors with outliers is 16 with the minimum percentage of 0.69%, maximum percentage of 33.09%, average percentage of 4.19%, and standard deviation percentage of 8.82%.
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 10, 'n_estimators': 50} |
1031-15-1-1-2-classification.csv | A multivariate classification time-series dataset consists of 7593 samples and 15 features with 15 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.37 showing a Unbalanced dataset. Among the 7593 samples the target ground-truth class has changed 1307 times representing a percentage of 17.29%. There are 15 features in the dataset
Among the numerical predictors, the series has 15 numerical features detected as Stationary out of the 15 numerical features using the dickey-fuller test and the rest are Unstationary. 13 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.038. to 1.512 and kurtosis values of 0.41 to 6.09. The fractal dimension analysis yields values ranging from -0.58 to -0.11 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.95, maximum 1.00, mean 0.11, and standard deviation 0.51. The count of numerical predictors with outliers is 15 with the minimum percentage of 12.75%, maximum percentage of 36.45%, average percentage of 31.06%, and standard deviation percentage of 6.92%.
The dataset is converted into a simple classification task by extracting the previously described features. | ElasticNetClassifier | {'C': 100.0, 'l1_ratio': 0.001, 'penalty': 'elasticnet', 'solver': 'saga'} |
1031-5-2-1-4-classification.csv | A multivariate classification time-series dataset consists of 7549 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.45 showing a Unbalanced dataset. Among the 7549 samples the target ground-truth class has changed 1401 times representing a percentage of 18.64%. There are 16 features in the dataset
Among the numerical predictors, the series has 13 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 16 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.114. to 20.833 and kurtosis values of 0.00 to 600.98. The fractal dimension analysis yields values ranging from -0.65 to -0.11 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.88, maximum 1.00, mean 0.08, and standard deviation 0.52. The count of numerical predictors with outliers is 16 with the minimum percentage of 0.07%, maximum percentage of 18.30%, average percentage of 7.03%, and standard deviation percentage of 5.27%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 1.0, 'n_estimators': 50} |
1031-12-1-1-4-classification.csv | A multivariate classification time-series dataset consists of 1198 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.35 showing a Unbalanced dataset. Among the 1198 samples the target ground-truth class has changed 181 times representing a percentage of 15.52%. There are 16 features in the dataset
Among the numerical predictors, the series has 13 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 15 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.032. to 10.809 and kurtosis values of 0.67 to 129.92. The fractal dimension analysis yields values ranging from -0.57 to -0.13 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.96, maximum 1.00, mean 0.11, and standard deviation 0.46. The count of numerical predictors with outliers is 16 with the minimum percentage of 2.92%, maximum percentage of 45.97%, average percentage of 40.05%, and standard deviation percentage of 10.94%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=5), 'learning_rate': 0.1, 'n_estimators': 250} |
1031-11-3-1-4-classification.csv | A multivariate classification time-series dataset consists of 6259 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.35 showing a Unbalanced dataset. Among the 6259 samples the target ground-truth class has changed 286 times representing a percentage of 4.59%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 16 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.205. to 13.261 and kurtosis values of 7.39 to 232.89. The fractal dimension analysis yields values ranging from -0.59 to -0.15 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.96, maximum 1.00, mean 0.11, and standard deviation 0.45. The count of numerical predictors with outliers is 16 with the minimum percentage of 13.78%, maximum percentage of 13.78%, average percentage of 13.78%, and standard deviation percentage of 0.00%.
The dataset is converted into a simple classification task by extracting the previously described features. | AdaboostClassifier | {'estimator': DecisionTreeClassifier(max_depth=2), 'learning_rate': 0.1, 'n_estimators': 50} |
1031-39-2-1-5-classification.csv | A multivariate classification time-series dataset consists of 6192 samples and 16 features with 15 numerical and 1 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0
The target column has 3 classes with entropy value 1.45 showing a Unbalanced dataset. Among the 6192 samples the target ground-truth class has changed 1215 times representing a percentage of 19.73%. There are 16 features in the dataset with a ratio of numerical to categorical features of 15.0.
Among the numerical predictors, the series has 12 numerical features detected as Stationary out of the 15 numerical features using the dickey-fuller test and the rest are Unstationary. 15 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.075. to 1.213 and kurtosis values of 0.08 to 3.36. The fractal dimension analysis yields values ranging from -0.59 to -0.13 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.78, maximum 1.00, mean 0.10, and standard deviation 0.56. The count of numerical predictors with outliers is 15 with the minimum percentage of 0.10%, maximum percentage of 24.57%, average percentage of 10.11%, and standard deviation percentage of 6.41%.
Among the categorical predictors, the count of symbols ranges from 76 to 76 with a minimum entropy value 0.357095452878317, maximum entropy 0.357095452878317, mean 0.357095452878317, and standard deviation 0.0,
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
1020-59-3-classification.csv | A multivariate classification time-series dataset consists of 7012 samples and 11 features with 9 numerical and 2 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00. Similarly, the missing values percentages for categorical features range from 0 to 0 with mean 0.0 and standard deviation 0.0.
The target column has 3 classes with entropy value 1.38 showing a Unbalanced dataset. Among the 7012 samples the target ground-truth class has changed 765 times representing a percentage of 10.96%. There are 11 features in the dataset with a ratio of numerical to categorical features of 4.5.
Among the numerical predictors, the series has 8 numerical features detected as Stationary out of the 9 numerical features using the dickey-fuller test and the rest are Unstationary. 9 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 38 seasonality components detected in the numerical predictors. The top 3 common seasonality components are represented using sinusoidal waves. of periods 1402,1753,2337 The numerical predictors also exhibit skewness values ranging from 0.303. to 4.874 and kurtosis values of 0.25 to 54.36. The fractal dimension analysis yields values ranging from -0.66 to -0.14 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.81, maximum 1.00, mean 0.14, and standard deviation 0.47. The count of numerical predictors with outliers is 7 with the minimum percentage of 0.00%, maximum percentage of 8.38%, average percentage of 3.21%, and standard deviation percentage of 2.84%.
Among the categorical predictors, the count of symbols ranges from 17 to 65 with a minimum entropy value 0.604688584813934, maximum entropy 3.9356770706012485, mean 2.270182827707591, and standard deviation 1.6654942428936572,
The dataset is converted into a simple classification task by extracting the previously described features. | RandomForestClassifier | {'max_depth': 10, 'n_estimators': 30} |
1031-26-1-1-2-classification.csv | A multivariate classification time-series dataset consists of 7593 samples and 16 features with 16 numerical and 0 categorical features. Each instance has a window length of 24. The dataset has a sampling rate of 60.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 3 classes with entropy value 1.35 showing a Unbalanced dataset. Among the 7593 samples the target ground-truth class has changed 1296 times representing a percentage of 17.15%. There are 16 features in the dataset
Among the numerical predictors, the series has 16 numerical features detected as Stationary out of the 16 numerical features using the dickey-fuller test and the rest are Unstationary. 14 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.023. to 19.258 and kurtosis values of 0.40 to 514.95. The fractal dimension analysis yields values ranging from -0.60 to -0.12 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.89, maximum 1.00, mean 0.11, and standard deviation 0.50. The count of numerical predictors with outliers is 15 with the minimum percentage of 0.00%, maximum percentage of 15.70%, average percentage of 7.60%, and standard deviation percentage of 3.65%.
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
1026-classification.csv | A multivariate classification time-series dataset consists of 1460 samples and 5 features with 5 numerical and 0 categorical features. Each instance has a window length of 7. The dataset has a sampling rate of 1440.0 minutes. The dataset has a missing values percentage of 0.0%. The missing values percentages for numerical features range from 0 to 0 with mean 0.00 and standard deviation 0.00.
The target column has 2 classes with entropy value 0.99 showing a Balanced dataset. Among the 1460 samples the target ground-truth class has changed 7 times representing a percentage of 0.48%. There are 5 features in the dataset
Among the numerical predictors, the series has 1 numerical features detected as Stationary out of the 5 numerical features using the dickey-fuller test and the rest are Unstationary. 5 of them are Multiplicative time-series features and the rest are Additive time-series features. There is an average of 0 seasonality components detected in the numerical predictors. The top 0 common seasonality components are represented using sinusoidal waves. The numerical predictors also exhibit skewness values ranging from 0.230. to 1.325 and kurtosis values of 1.11 to 4.22. The fractal dimension analysis yields values ranging from -0.44 to -0.31 indicating a Complex and Irregular time-series structure for the numerical predictors. The correlation values among the numerical predictors have a minimum of -0.23, maximum 1.00, mean 0.61, and standard deviation 0.57. The count of numerical predictors with outliers is 1 with the minimum percentage of 0.00%, maximum percentage of 9.13%, average percentage of 1.83%, and standard deviation percentage of 4.09%.
The dataset is converted into a simple classification task by extracting the previously described features. | XGBoostClassifier | {'learning_rate': 0.1, 'max_depth': 5, 'n_estimators': 20, 'reg_lambda': 0.2} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.