Predicting the Geoeffectiveness of CMEs Using Machine Learning

Coronal mass ejections (CMEs) are the most geoeffective space weather phenomena, being associated with large geomagnetic storms, having potential to cause disturbances telecommunication, satellite network disruptions, power grid damages and failures. Thus, considering these storms' effects on human activities, accurate forecasts of geoeffectiveness CMEs paramount. This work focuses experimenting different machine learning methods trained white-light coronagraph datasets close sun CMEs, estimate whether such a newly erupting ejection has induce activity. We developed binary classification models using logistic regression, K-Nearest Neighbors, Support Vector Machines, feed forward artificial neural networks, as well ensemble models. At this time, we limited our forecast exclusively use solar onset parameters, ensure extended warning times. discuss main challenges task, namely extreme imbalance between number ineffective events in dataset, along their numerous similarities available variables. show that even conditions, adequate hit rates can be achieved