Solar Flare Prediction Using a Hybrid Convolutional Neural Network and Transformer Model

This paper presents a hybrid convolutional neural network and transformer model for predicting solar flare intensity. Solar flares are categorized into five classes (A, B, C, M, X), with M and X classes capable of causing significant disruptions to communication systems, satellites, and power grids on Earth. The model leverages the physical mechanism of sudden magnetic energy release in the Sun's atmosphere to improve prediction accuracy across different flare classes.