Development of NARX Neural Network based Optimized Deep Learning Algorithm for Adaptive Rainfall Prediction in North-Western Himalayas

Abstract

The paper proposes a Nonlinear Auto-Regressive with external input (NARX) neural network-based optimized deep learning algorithm for adaptive rainfall prediction in the North-Western Himalayas. The study focuses on examining the impact of the size of the training dataset and the number of neurons on the performance of the proposed algorithm in predicting monthly rainfall. This study presents a adaptive rainfall prediction model based on NARX Neural Network Model and Levenberg-Marquardt (LM) optimization algorithm to forecast the Indian summer monsoon rainfall (ISMR) using monthly seasonal time series data for Jammu &Kashmir and Indian rainfall data. The study finds that the algorithm's performance improves with an increase in the size of the training dataset and the number of neurons. Mean Squared Error (MSE) and Percentage Root Mean Square Difference (PRD) metric estimators are used to evaluate the performance of the proposed model with varying number of hidden neuron and varying training data size. The authors conclude that the proposed algorithm can be an effective tool for predicting rainfall in the North-Western Himalayas region and can aid in disaster preparedness and flood management. However, the study's limitation is that it focuses only on the North-Western Himalayas region, and further research is required to generalize the results to other regions. Overall, this study makes a valuable contribution to the field of rainfall prediction by showcasing the potential of NARX neural network-based optimized deep learning algorithms in predicting complex meteorological phenomena.