MagicTelescope: Particle Detection Data for Gamma/Hadron Classification

MagicTelescope is a dataset from high-energy physics for classifying gamma rays and hadrons. This is a 2000-row, 100-column subsample with 10 target classes, created by Eddie Bergman from the original OpenML dataset. The license is us-pd, placing it in the public domain.

Use Cases

• Benchmarking multi-class classification models on physics data.
• Developing feature selection techniques on a dataset with 100 columns.
• Training models to distinguish between gamma-ray and hadron signals from telescope observations.

Strengths

• Subsampling parameters are explicitly defined: seed=1, nrows=2000, ncols=100, nclasses=10.
• The original dataset is a known benchmark (OpenML ID 44125) for particle detection.
• Creation method is transparent, with provided source code for stratified subsampling.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the full original dataset is unknown, limiting context for this subsample.
• Last update date is unknown; freshness unverified.

Provenance

Source

OpenML dataset 'MagicTelescope' (ID 44125), subsampled by Eddie Bergman.

Collection Method

Algorithmic subsampling with stratification, as described in the provided Python code.