This project focuses on building and evaluating various classifier models to predict rainfall using a noisy, imbalanced, and often inaccurate weather dataset. The data, collected from multiple weather stations, includes features such as rainfall amounts, wind trajectory, and other relevant weather parameters.
The primary objective is to predict whether it will rain tomorrow based on historical weather data. Through model selection, feature engineering, and handling of the dataset’s inherent challenges, we aim to provide accurate forecasts despite the data’s imperfections.