Enhancing Wheat – Yield Forecasting with Ensemble Random Forest Machine Learning Technique

Abstract

Economy and food security of our country depends heavily on wheat, therefore precise yield forecasts is crucial for resource management and planning. Although traditional methods, such remote sensing and manual field surveys have been widely employed, it is still unclear how well they capture yield variance throughout various growth phases. Predicting crop yields is essential for increasing productivity, controlling risks, guaranteeing food security, and boosting agriculture's general sustainability. The Random Forest (RF) machine learning approach was assessed in this study for its capacity to forecast wheat production of the country in the period 2023-2030, responses to soil and climate-related parameters with Multiple Linear Regression (MLR). Evaluation measures included Error in Root Mean Square (RMSE) also Absolute Mean Error (AME), with RF. This finding indicates, RF is better than Existing Markov Chain method at predicting wheat production.