Rainfall Prediction Using Machine Learning

Abstract

Rainfall prediction is a crucial task with significant importance in agriculture, water resource management, and disaster prevention. Accurate prediction of rainfall helps farmers plan crop activities, supports efficient water utilization, and aids authorities in preparing for extreme weather conditions. This project focuses on developing a machine learning-based system to predict the daily occurrence of rainfall (yes or no) using historical meteorological data. The proposed approach utilizes a Logistic Regression classifier trained on a dataset containing key weather-related features such as pressure, temperature, humidity, dew point, wind speed, and cloud cover. The methodology follows a structured pipeline, including data preprocessing, exploratory data analysis using correlation heatmaps, feature scaling, model training, and evaluation. The model achieved an accuracy of approximately 76.7% on the test dataset, demonstrating its ability to effectively capture patterns in weather data. Further analysis of model coefficients revealed that cloud cover and humidity are strong positive indicators of rainfall, while minimum temperature and sunshine show a negative influence.