Rice and Potato Yield Prediction Using Artificial Intelligence Techniques

Crop yield prediction during the growing season is important for crop income, insurance projections and even ensuring food security. Yet, modeling crop yield is challenging because of the complexity of the relationships between crop growth and the interrelated predictor variables. This research work employed artificial intelligence (AI) technique for rice and potato crop yield prediction model in the region of Tarakeswar block, Hooghly District, West Bengal, for rice and potato. The major variables used were climatic factors, static soil parameters, available soil nutrient, agricultural practice parameters, farm mechanization, terrain distribution and socioeconomic condition. The analyzed datasets covered 2017 to 2018 seasons and were split into two parts with seventy percent data used for model training and the remaining thirty percent for validation. The mean rice and potato yield obtained from the seventy-farm plot location was about 4.68 t/ha and 18.67 t/ ha, whereas the artificial neural networks (ANN) model estimated with 97% accuracy and R2 value of both the crop is 0.93 and 0.94 with an RMSE of 0.29 t/ha and 1.34 t/ha, respectively. Deep neural networks (DNN) outperformed among the three model, where only support vector machine (SVM) had a sound performance for the training data but low for the validation dataset due to overfitting problem within RMSE and R2 value. The optimized DNN model produced the highest prediction accuracy 98% for rice and potato crop (RMSE = 0.20 ton/ha and 0.95 t/ ha; R2 = 0.98 and 0.97, respectively), which indicates good correlation between the field-measured crop yield and estimated yield. These adopted methodology for prediction crop yield to provide recommendation to the farmers, decision makers and stakeholders can make farming more efficient and profitable. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2022.