Indian Journal of Science and Technology

Abstract

Objectives: The main goal of this research is to design and analyze the performance of intelligent Machine Learning (ML) algorithms such as Support Vector Machines and Naïve Bayes for the prediction of rainfall in the three districts of Jalgaon, Dhule, and Nandurbar in Maharashtra, India. This research attempts to identify the factors that contribute to the occurrence of rainfall at the regional level. Methods: The data used in this study are meteorological variable parameters from the previous ten years for the 21 locations in the study area. The predictive performance of the model was validated using several statistical metrics such as precision, accuracy, and f-score. Findings: The experimental results demonstrate that both Machine Learning models provided acceptable predictions of rainfall. However, the Support Vector Machine was found to be the best model, with 93% accuracy, for predicting rainfall. In large datasets, prediction results improve significantly. Novelty: The Support Vector Machine Model outperforms the Naïve Bayes and Decision Tree models. The study investigates how the meteorological parameters of an atmospheric pressure of 1008 Mb, with winds flowing from the western ghats from the west, and temperatures ranging from 20 to 30 oC affect the decisionmaking capability of the Support Vector Machine model for drawing the precise hyperplane. Further, Identifying the rain and non-rain situations precisely. Keywords: Rainfall Prediction; Machine Learning; Support Vector Machines; Naïve Bayes; Meteorological Parameters