Solar Flare Prediction from Time Series Data 2020
A Track in the IEEE Big Data 2020 Big Data Cup Challenge
Description of the Data
The magnetic field parameters used in this work are described in [Solar flare prediction using SDO/HMI vector magnetic field data with a machine-learning algorithm] and are listed in the table below. Those parameter names listed in red are described in the above document, but are not available in the SHARPs data accessible at JSOC. Our dataset also has the addition of the max x-ray luminosity observed over +/- six minutes around the representative time step. This is used because the reporting cadence from the GOES satellite is different than that of the source for the other parameters.
In addition to the parameters listed above, we have introduced a weighting of each parameter based on their distance from the polarity inversion line on the active region the parameters were calculated on. These parameters are experimental, and we have not evaluated their usefulness, so use at your own discretion. The naming of these parameters in the file are the same as the other parameters without weighting, but are preceded by the PIR_ designation.
More details of the file format, can be found on the competition Kaggle data page.