Brain Emotional Learning-based Prediction Model For the Prediction of Geomagnetic Storms

TL;DR

This paper introduces BELPM, a novel brain emotional learning-inspired model for predicting geomagnetic storms using AE and Dst indices, demonstrating competitive accuracy compared to existing methods.

Contribution

The paper presents a new BELPM model based on emotional system regions for geomagnetic storm prediction, integrating adaptive networks and specific learning algorithms.

Findings

BELPM achieves reasonable accuracy in short-term and long-term predictions.

BELPM outperforms ANFIS and WKNN in geomagnetic storm prediction.

The model effectively utilizes AE and Dst indices for forecasting.

Abstract

This study suggests a new data-driven model for the prediction of geomagnetic storm. The model which is an instance of Brain Emotional Learning Inspired Models (BELIMs), is known as the Brain Emotional Learning-based Prediction Model (BELPM). BELPM consists of four main subsystems; the connection between these subsystems has been mimicked by the corresponding regions of the emotional system. The functions of these subsystems are explained using adaptive networks. The learning algorithm of BELPM is defined using the steepest descent (SD) and the least square estimator (LSE). BELPM is employed to predict geomagnetic storms using two geomagnetic indices, Auroral Electrojet (AE) Index and Disturbance Time (Dst) Index. To evaluate the performance of BELPM, the obtained results have been compared with ANFIS, WKNN and other instances of BELIMs. The results verify that BELPM has the capability to achieve a reasonable accuracy for both the short-term and the long-term geomagnetic storms prediction.