Title: Enhancing rainfall prediction at Dapoli, Ratnagiri District, India: a unified approach with multilayer perceptron modelling

Authors: Krishna Gopal Dhal; Manab Nandi; Arunita Das; Arpan Garai

Addresses: Department of Computer Science, Midnapore College (Autonomous), Paschim Medinipur, West Bengal, India ' Department of Computer Application, Midnapore College (Autonomous), Paschim Medinipur, West Bengal, India ' Department of Computer Science, Midnapore College (Autonomous), Paschim Medinipur, West Bengal, India ' Department of Computer Science and Engineering, Vellore Institute of Technology, Vellore, Tamil Nadu, India

Abstract: Accurate prediction of rainfall is critically challenging due to its erratic nature. Prediction of rainfall plays a major role in various sectors such as agriculture, water management, disaster preparedness, etc., especially in vulnerable areas like Dapoli, India. The presence of generalisations of a prediction model is crucial, as it showcases the applicability and usefulness of the model in diverse settings. This research paper addresses the enhancement of prior existing rainfall predictive models for Dapoli by proposing a single multilayer perceptron (MLP) neural network. This model is capable of generalising across different datasets spanning 1983-2016, 1983-1992, 1993-2002 and 2003-2016, unlike the existing works where separate models are used. Despite having slightly lower correlation coefficients, this single model achieves a lower range of RMSE values (ranging from 10.734 to 16.185) compared to the existing four models (ranging from 8.14228 to 22.05690) after performing extensive experiments by tuning various hyperparameters and data scaling processes. Therefore, this research paper unlocks a more efficient and generalised rainfall prediction approach across diverse datasets for Dapoli by showcasing the potential of MLP networks for complex weather prediction.

Keywords: rainfall prediction; artificial neural networks; ANN; multilayer perceptron; MLP.

DOI: 10.1504/IJAISC.2025.151137

International Journal of Artificial Intelligence and Soft Computing, 2025 Vol.9 No.3, pp.251 - 282

Received: 22 Mar 2025
Accepted: 11 Sep 2025
Published online: 14 Jan 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article