Machine Learning Approaches to Predict Rice Yield of Bangladesh

Rice is one of the most grown and consumed crops of the world which largely depends on the weather. To assure global food security, it is extremely important to predict rice yield production for the near future. Machine Learning (ML) can be used to predict or estimate values of different parameters almost accurately. Agriculture and Weather Forecast are two sectors where ML techniques are used extensively. In this paper, we have used weather forecast scenarios to predict rice production in the region of Bangladesh. Data from different sources are integrated to predict the yield production of rice which is considered as the target prediction for this experiment. After predicting the weather parameter values of a particular time frame, we have used the values for target prediction. Several ML-based regression models such as Linear, Polynomial, Support Vector Regression are implemented and results are collected according to the accuracy for each parameter of weather. Finally, Support Vector Regression (SVR) is used for target prediction and achieves 95.89% accuracy to predict the rice yield production amount.