Seismic-Bumps: High-Energy Seismic Hazard Forecasts from a Polish Coal Mine

Seismic-bumps is a dataset of 2,584 instances describing the problem of forecasting high-energy seismic bumps in a coal mine. The data come from two longwalls in a Polish coal mine and were published by researchers from the Silesian University of Technology and the EMAG Institute of Innovative Technologies. It contains 18 attributes and a binary class label indicating a hazardous or non-hazardous state.

Use Cases

• Predict hazardous seismic states based on seismic and seismoacoustic hazard assessment results.
• Analyze the relationship between seismic energy, pulse counts, and their deviations from historical averages.
• Classify shift safety conditions using features like shift type and the number of recorded seismic bumps.
• Build binary classification models for mine safety using the provided class labels for hazardous and non-hazardous states.

Strengths

• Contains 2,584 complete instances with no missing attribute values.
• Includes 18 attributes covering multiple seismic monitoring methods and energy measurements.
• Class distribution is explicitly defined, with 170 hazardous and 2,414 non-hazardous instances.

Limitations

• Column-level documentation is incomplete; the full list of 18 attributes is not detailed in the provided input.
• Last update date is unknown; freshness unverified.
• Data may reflect geographic and operational bias inherent to the specific Polish coal mine source.

Provenance

Source

Marek Sikora and Lukasz Wrobel, Silesian University of Technology and EMAG Institute of Innovative Technologies.

Collection Method

Collected by seismic hazard monitoring systems in a coal mine.

Geography

Polish coal mine (specific mine not named).