
This release contains the source code and implementation of a machine learning-based model for predicting annual coffee yield in the Western Hararghe Zone, Ethiopia. The project uses multi-year agronomic, soil, and climatic data collected from five districts and applies several supervised learning algorithms, including Ridge Regression, ElasticNet, Random Forest, XGBoost, Gradient Boosting, and Support Vector Regression. Hyperparameter tuning was performed using GridSearchCV to improve model performance and ensure robust prediction accuracy. The final results show that Ridge Regression achieved the best performance in terms of accuracy and stability. This repository also includes data preprocessing steps, model training, evaluation metrics, and visualization outputs to support reproducibility and further research in agricultural forecasting and climate-smart farming systems.