Comparative Analysis of Statistical and Machine Learning Techniques for Rice Yield Forecasting for Chhattisgarh, India

Crop yield forecasting before harvesting is critical for the creation, implementation, and optimization of policies related to food safety as well agro-product storage marketing. growth development are influenced by weather. Therefore, models using weather variables can provide reliable predictions crop yields. It be tough select best production model. in this study, five alternative models, viz., stepwise multiple linear regression (SMLR), an artificial neural network (ANN), least absolute shrinkage selection operator (LASSO), elastic net (ELNET), ridge regression, were compared order discover model rice prediction. The outputs from individual used build ensemble generalized (GLM), random forest (RF), cubist ELNET methods. For previous 21 years, historical statistics meteorological data collected three districts under separate agro-climatic zones Chhattisgarh, Raipur Chhattisgarh plains, Surguja northern hills, Bastar southern plateau. calibrated 80% these datasets, remaining 20% was validation models. present study concluded that forecasting, performance ANN good (Rcal2 = 1, Rval2= 1 RMSEcal 0.002, RMSEval 0.003) 0.99 0.004, 0.214) other whereas Bastar, 90, 0.48) LASSO 93, 0.568) performed better. better Surguja, all methods comparable, (RF) better, with R2 0.85 0.81 calibration validation, respectively, GLM, cubist, approach.