Datasets:
Formats:
parquet
Size:
100M - 1B
ArXiv:
Tags:
timeseries
timeseries clustering
changepoint-detection
correlation-structure
Synthetic
benchmark
DOI:
License:
Correct indent for subtitle
Browse files
README.md
CHANGED
|
@@ -1388,7 +1388,7 @@ and messy real-world data.
|
|
| 1388 |
- Investigating how **data preprocessing** affects correlation structure discovery
|
| 1389 |
- Establishing **performance thresholds** for high-quality clustering result
|
| 1390 |
|
| 1391 |
-
|
| 1392 |
The dataset features 23 distinct correlation structures representing different combinations of strong positive, negligible, and strong negative correlations between three time series variates. These structures are based on meaningful thresholds for strong negative ([-1,-0.7]), negligible ([-0.2,0.2]), and strong positive ([0.7,1]) correlations.
|
| 1393 |
|
| 1394 |
### Dataset Structure
|
|
|
|
| 1388 |
- Investigating how **data preprocessing** affects correlation structure discovery
|
| 1389 |
- Establishing **performance thresholds** for high-quality clustering result
|
| 1390 |
|
| 1391 |
+
### Correlation Structures
|
| 1392 |
The dataset features 23 distinct correlation structures representing different combinations of strong positive, negligible, and strong negative correlations between three time series variates. These structures are based on meaningful thresholds for strong negative ([-1,-0.7]), negligible ([-0.2,0.2]), and strong positive ([0.7,1]) correlations.
|
| 1393 |
|
| 1394 |
### Dataset Structure
|