Modeling Dst with Recurrent EM Neural Networks

Recurrent Neural Networks have been used extensively for space weather forecasts of geomagnetospheric disturbances. One of the major drawbacks for reliable forecasts have been the use of training algorithms that are unable to account for model uncertainty and noise in data. We propose a probabilistic training algorithm based on the Expectation Maximization framework for parameterization of the model, which makes use of a forward filtering and backward smoothing Expectation step, and a Maximization step in which the model uncertainty and measurement noise estimates are computed. The inputs to the network are based on three parameters of the interplanetary magnetic field (IMF ), b z , b 2 , and b y 2 , along with the D st index. Through numerical experimentation it is shown that the proposed model allows for reliable forecasts and also outperforms other neural time series models trained with the Extended Kalman Filter, and gradient descent learning.