license: cc0-1.0
Dataset Card for Iris Dataset
Dataset Summary
The Iris dataset is a classic benchmark dataset in machine learning, introduced by Ronald Fisher in 1936. It contains measurements of 150 iris flowers from three species: Iris setosa, Iris versicolor, and Iris virginica. Each sample is described by four features: sepal length, sepal width, petal length, and petal width, all measured in centimeters. The dataset is widely used for classification tasks and data visualization practices due to its simplicity and balanced structure.
Supported Tasks and Leaderboards
The Iris dataset supports the following tasks:
- Classification: Predict the species of an iris flower based on its features.
- Visualization: Explore relationships between features using scatter plots and other methods.
No active leaderboard is associated with this dataset.
Languages
The dataset does not contain textual data; it consists entirely of numerical measurements and categorical labels.
Dataset Structure
Data Instances
Each instance represents an iris flower with the following attributes:
- Sepal length (float)
- Sepal width (float)
- Petal length (float)
- Petal width (float)
- Species (categorical: 0 = Iris setosa, 1 = Iris versicolor, 2 = Iris virginica)
Data Splits
The dataset consists of 150 samples with no predefined splits. Users can create custom splits for training, validation, and testing.
Dataset Creation
The dataset was originally collected by Edgar Anderson and introduced by Ronald Fisher to demonstrate linear discriminant analysis. It has since become a standard benchmark for testing classification algorithms.
Considerations for Using the Data
Social Impact of Dataset
The Iris dataset is primarily used for educational purposes and algorithm benchmarking. It does not contain sensitive or personal information.
Limitations
The dataset is small in size (150 samples) and may not represent real-world complexities in classification tasks. Additionally, its simplicity may limit its applicability to advanced machine learning models.
Additional Information
Homepage
Not applicable.
Repository
Available through scikit-learn.
Paper
Fisher, R.A. (1936). "The use of multiple measurements in taxonomic problems."
Citation
If you use this dataset, please cite Fisher's original paper: