Machine Learning Based Prediction of Agricultural Foodgrain Production in Indian States Using Agricultural, Irrigation and Fertilizer Indicators

Abstract

Agricultural foodgrain production is influenced by multiple interacting factors, including cultivated area, irrigation availability, fertilizer use, rainfall, and previous production levels. Reliable prediction of foodgrain production can support agricultural planning, resource allocation, and food-security management. This study develops a machine-learning-based framework for predicting agricultural foodgrain production across Indian states using agricultural, irrigation, fertilizer, climatic, and historical production indicators. A state-year dataset comprising 600 observations from 25 Indian states during 2000–01 to 2023–24 was developed. Seven predictor variables, namely Net Sown Area, Gross Sown Area, Net Irrigated Area, Gross Irrigated Area, Fertilizer Consumption, Annual Rainfall, and Previous-Year Foodgrain Production, were used to predict Foodgrain Production. Two ensemble regression algorithms, Random Forest (RF) and Extreme Gradient Boosting (XGBoost), were developed and evaluated using the correlation coefficient (R), coefficient of determination (R²), root mean square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and scatter index (SI). On the testing dataset, RF achieved R = 0.996298 and R² = 0.987867, with RMSE = 2003.8967 thousand tonnes, MAE = 957.6846 thousand tonnes, MAPE = 6.919%, and SI = 0.148889. XGBoost achieved R = 0.996503 and R² = 0.983717, with RMSE = 2321.4478 thousand tonnes, MAE = 933.7023 thousand tonnes, MAPE = 11.312%, and SI = 0.172483. Based on the overall testing performance, RF was identified as the better-performing model. Feature-importance analysis further identified Previous-Year Foodgrain Production as the most influential predictor in both RF and XGBoost. The findings demonstrate the potential of ensemble machine-learning models, particularly RF, for data-driven prediction of agricultural foodgrain production across Indian states.