Electricity: A Stratified Subsampled Classification Dataset

A 2000-row, 100-column subset of the OpenML electricity dataset, generated with a random seed of 4. It was created by Eddie Bergman using a Python script to uniformly sample columns and rows while preserving up to 10 target classes. The dataset is derived from the original OpenML dataset with ID 44156.

Use Cases

• Benchmarking classification model performance on a reduced-scale electricity dataset.
• Testing feature selection methods on a dataset with 100 sampled columns.
• Evaluating model behavior on stratified data with up to 10 target classes.
• Comparing algorithm results against a known subsampling seed for reproducibility.

Strengths

• Controlled subsampling with a specified random seed (4) for reproducibility.
• Stratified sampling was applied, likely preserving class distribution.
• Dataset size is explicitly defined as 2000 rows and 100 columns.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Description metadata is limited; actual data quality requires manual inspection after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

OpenML (dataset ID 44156)

Collection Method

Subsampled via a custom Python script using numpy.random and train_test_split.