P. Bhandari

Comparative Benchmarking Study of Deep Learning Models for Geomagnetic Storms Prediction

The results have shown that it is possible to model nonlinear temporal dependencies in geomagnetic observations using recurrent architectures, and the use of traditional baselines along with repeated-seeds and ablation tests makes the evaluation more thorough.