Yield Prediction of Rabi Crops in India Using XGBoost Regressor

Agriculture is at the heart of India's economy, with two key cropping seasons: Kharif, grown in the summer, and Rabi, grown in the winter. Rabi crops like wheat, gram, and mustard are cultivated from October to April, particularly in Northwest India. Accurately forecasting the yield of these crops is vital for efficient planning, resource management, and ensuring food security, especially given the unpredictability of climate patterns. In this study, we predict yield based on rainfall and various climatic factors which impact Rabi crop. Using data from different states on crop production and rainfall, we built predictive models to estimate yields. After processing and refining the data, we utilized advanced machine learning methods like Random Forest, XGBoost, and LightGBM to assess how rainfall affects crop performance. Additionally, we improved the accuracy of the Random Forest model through hyperparameter tuning. The models' accuracy was evaluated using metrics like Mean Squared Error (MSE) and R-squared (R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>), with cross-validation ensuring consistent results. Of the models tested, Random Forest, XGBoost, and LightGBM performed best, with the latter two offering the most precise predictions. The results show that XGBoost achieved the best accuracy with a maximum of 96% and after hyperparameter tuning showed a better result of 97%. In this paper we highlight how integrating cutting-edge machine learning algorithms can improve the precision of crop yield forecasts, aiding agricultural planning and helping to secure food supplies for the future.