Comparing Different Machine Learning and Deep Learning Models for Daily Rainfall Prediction at Kerala Point Location

Accurate prediction of daily rainfall is crucial for water resource management, disaster preparedness, and agricultural planning in monsoon-prone regions such as Kerala. Traditional statistical models cannot capture the nonlinear nature of the rainfall patterns. This paper attempts to compare various Machine Learning (ML) and Deep Learning (DL) models, including Long Short-Term Memory (LSTM), XGBoost, LightGBM, and Gradient Boosting, to assess the performance of daily rainfall forecasting at a Kerala point location. Using historical rainfall data, from 1981 to 2023, The models were trained, tested, and evaluated for performance based on the Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>. Results: Deep learning approaches especially LSTM perform very well and outshine the traditional approaches and capture temporal dependencies Also the ensemble approach has a maximum predictive accuracy, thus the hybrid approach can be useful for rainfall time series forecasting. This research focuses on the benefits of advanced learning techniques in weather prediction and gives suggestions for future improvements, including real-time meteorological data integration.