Smart harvest: a multicrop, rainfall-integrated model comparison framework for yield prediction

Abstract

This work introduces Smart Harvest, a predictive analytics model for predicting rice, wheat, jowar, and bajra crop yields. The model investigates the rainfall impact on yield through the use of historical production and climatic data. ARIMA, ARIMAX, SVR, ANN, and Random Forest models are considered and compared for performance and accuracy. The findings indicate the robustness of rainfall-integrated models in enhancing prediction accuracy. This method facilitates data-driven farm planning and risk management.