Description:

Machine learning creates new opportunities in the Space Weather community for identification, classification, modeling and forecasting. Bridging the gap between the space science and the machine learning community is crucial to working with the enormous datasets collected by space missions. Large, and freely available datasets of in-situ and remote observations collected over several decades of space missions allow for space weather to be an ideal application for machine learning. Utilizing imagery, geomagnetic indices, particle fluxes, magnetograms, and more, one can understand more about complex nature of the solar-terrestiral system in which we live. This session welcomes presentations on the advances in space weather utilizing information theory, neural networks, clustering algorithms, nonlinear auto-regression models, and other nontraditional approaches that take into account the nonlinear and complex dynamics of space weather to improve forecasting, predictions, classification, identification, and uncertainty propagation.