CME Arrival Time Prediction with Artificial Intelligence via Hybrid Features

Predicting the accuracy of Coronal Mass Ejections (CMEs) arrival times is crucial for mitigating the adverse effects of space weather on technological infrastructure. This research introduces an innovative approach to CME arrival time prediction by using artificial intelligence (AI) models enhanced with hybrid features. Hybrid features combine CME parameters extracted from SOHO LASCO C2 observations such as CME speed, angular width, and position angle with solar wind parameters such as Bx, By, Bz, plasma density, plasma temperature, flow speed, and B Field Magnitude. Leveraging advanced machine learning algorithms, our approach significantly improves prediction accuracy compared to other models. We achieve superior results with a mean absolute error (MAE) of only 6.89 hours and a root mean square error (RMSE) of 10.69 hours. This research presents the potential of hybrid features in advancing space weather forecasting capabilities and providing the accuracy of timely warnings of CME arrival to Earth to protect critical technological infrastructure.