Leveraging ARMA and ARMAX Time-Series Forecasting Models for Rainfall Prediction

Abstract

The downpour of rainwater in the world is growing increasingly variable, making forecasting more difficult. The Indian Meteorological Department (IMD) currently makes use of Hybrid and Mathematical approaches to forecasting Indian summer monsoon rainfall. As a corollary, analysts are barely able to predict the socioeconomic repercussions of floodwaters (excessive rain) or famine (fewer drops of rain). The amount of rain that falls relies on various factors, including the temperature levels in the atmosphere, wetness, wind, velocity, and wind conditions. Several methodologies have been used in weather prediction. In this paper, two forecasting approaches are preferred: ARMA (Auto-Regressive Moving Average) and ARMAX (Auto-Regressive Moving Average with exogenous variables). The Ensemble strategy effectiveness is evaluated using RMSE, MSE, and MAE. It is observed that the ARMA+ARMAX model has the least values of MAE at 4.921, MSE at 4.654and RMSE at 2.456 when compared with cutting-edge strategies.