February 23rd, 2018, 1:00 pm - 2:00 pm

February 23rd, 1:00 pm - 2:00 pm

FLARECAST: a computational/technological platform for solar flare prediction based on machine learning

Michele Piana (Dipartimento di Matematica Universita di Genova)

This talk will describe a computational and technological service for solar flare prediction developed within the framework of a EU project. The main aspects of this platform are that it exploits properties extracted from SDO/HMI magnetograms, it utilizes both supervised and unsupervised machine learning methods for prediction, and it is able to rank the properties on the basis of their impact on the prediction performance. Some final comments about possible multi-task characterization of the flaring event will be given.