ABSTRACT
This project leverages machine learning models and incorporates satellite data and remote sensing techniques to enhance rainfall prediction in Nigeria. By merging historical meteorological data with satellite imagery data, the performance of various models, including Linear Regression, LSTM, Random Forest, XGBoost, SVR, Lasso Regression, and ElasticNet Regression, is evaluated. Performance metrics such as Mean Squared Error (MSE), Mean Absolute Error (MAE), and Coefficient of Determination (R2) guide model assessments. The results highlight the strengths of ensemble techniques like Random Forest and XGBoost, as well as the temporal data capturing ability of LSTM networks. Future research directions are identified, emphasizing advanced ensemble methods, realtime data integration, and model customization by region. Challenges in accessing comprehensive meteorological data, particularly due to cost constraints, underscore the importance of open data initiatives. It aims to contribute to improved rainfall prediction in Nigeria, offering insights for future research, operational implementations, and to benefit agriculture, water resource management, and disaster preparedness in the country.
