Electricity Consumption: Subsampled Benchmark for ML with 10 Classes

A 2,000-row, 100-column subsample of a larger electricity dataset, created with a random seed of 1. It was generated by Eddie Bergman using a script that uniformly sampled rows and columns while preserving up to 10 target classes. The original dataset is licensed as a US public domain work.

Use Cases

• Benchmarking classification model performance on a structured, tabular dataset.
• Testing feature selection methods on a dataset with 100 sampled columns.
• Evaluating stratified sampling techniques for imbalanced class distributions.
• Comparing model scalability on a medium-sized dataset of 2,000 rows.

Strengths

• Controlled subsampling with a specific random seed (1) ensures reproducible benchmarking.
• Stratified sampling was applied, which likely preserves the original class distribution in the 2,000-row subset.
• The dataset has a defined structure with 100 features and up to 10 target classes for classification tasks.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the original source dataset is unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.

Provenance

Source

OpenML, derived from dataset ID 44156.

Collection Method

Algorithmic subsampling of a larger dataset using a custom Python script.