Kepler Stellar Flare Light Curves with Quasiperiodic Pulsation Detections

2,274 detrended stellar flare light curves derived from the Kepler flare catalogue, with a subset of 159 curves identified as containing quasiperiodic pulsations. The dataset was created by Sergey Belov and presented in 2024, based on the original 2015 catalogue. It is intended for time-series analysis and machine learning applications in astrophysics.

Use Cases

• Train a classifier to detect quasiperiodic pulsations (QPPs) in detrended flare light curves.
• Analyze the temporal characteristics of the 159 light curves with QPPs to model flare decay patterns.
• Apply time-series forecasting models to predict flare intensity based on the parabolic detrending method used.
• Benchmark fully convolutional network (FCN) architectures against the provided QPP detection results with 95% confidence.

Strengths

• Contains 2,274 curated and detrended stellar flare light curves.
• Includes a labeled subset of 159 light curves with high-confidence QPP detections.
• Data is derived from the authoritative Kepler flare catalogue.

Limitations

• The sample size of confirmed QPP events is relatively small (159 out of 2,274).
• Limited to flares observed by the Kepler mission, which may introduce observational biases.
• The raw, non-detrended light curves and specific columnar data are not described.

Provenance

Source

Kepler flare catalogue (Balona 2015), processed by Belov et al. (2024).

Collection Method

Original Kepler light curves were detrended using a parabolic fit; QPPs were detected using a fully convolutional network (FCN) approach.

Time Range

Covers the observational period of the Kepler mission, as reflected in the 2015 source catalogue.

Freshness

The dataset was last updated in April 2026, indicating recent maintenance.

Geography

Observations from the Kepler space telescope; specific stellar coordinates are not provided.