Ozone Level 8-Hour: Subsampled Classification Dataset with 10 Classes

A subsampled version of the ozone-level-8hr dataset from OpenML, generated with a random seed of 1. The dataset likely contains up to 2000 rows and 100 columns, with the target variable stratified across 10 classes. Eddie Bergman is listed as the author, and the data is released under a US public domain license.

Use Cases

• Training classification models to predict ozone levels based on atmospheric measurements.
• Benchmarking feature selection methods on a dataset with 100 potential predictor columns.
• Studying the performance of classifiers on a stratified subset of environmental data with 10 target classes.

Strengths

• The subsampling process explicitly targeted a maximum of 2000 rows and 100 columns, providing a controlled scale.
• The generation code indicates stratified sampling was used for the target variable across 10 classes, which can help maintain class distribution.

Limitations

• Row count is unknown; the actual size may be less than the 2000-row maximum specified in the generation parameters.
• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

OpenML (dataset ID 1487), subsampled by Eddie Bergman.

Collection Method

Algorithmic subsampling from a parent dataset using a specified random seed, column selection, and stratified row sampling.