ML and AI-Based Prediction of Precipitation Trends

Abstract

Rainfall patterns are vital in shaping ecosystems and supporting life on Earth. However, changing rainfall distribution and intensity due to environmental factors and climate change have become increasingly evident in recent years. This study aims to utilize machine learning (ML) and Artificial intelligence (AI) techniques to predict rainfall trends in the Kokrajhar region, a district in Assam, India. Accurate rainfall prediction is crucial for effective water resource management, agriculture planning, and disaster preparedness in the area. To achieve this objective, 31 years of historical rainfall data from Kokrajhar have been collected. Various machine learning algorithms, including Random Forests, MLP, ETS, and MEDIAN, have been employed. The models were trained and evaluated using the data set, with performance metrics mean absolute percentage error (MAPE). The results suggest that the ETS model (Average MAPE 424.11% after removing outliers) exhibited the best performance among the other models in the second fold. In the third fold, the MLP model outperformed the others. These findings demonstrate the effectiveness of AI and ML techniques in predicting rainfall trends in the Kokrajhar area.