Bangladeshi Paddy Yield Prediction Using Machine Learning

Abstract

Rice is the staple food for Bangladesh's population of 135 million. It provides two-thirds of the region's total calorie supply and half of its total protein consumption, supports nearly half of all rural jobs, and accounts for almost half of all land area. The rice industry accounts for 16% of Bangladesh's GDP and 50% of the agricultural GDP. Most of the country's 13 million farming households produce rice. The area used for growing rice, at over 10.5 million hectares, has been fairly stable over the previous three decades. In Asia, rice takes up around 75% of all farmland and 80% of all irrigated land. Accordingly, rice is extremely important to the diets of the people of Bangladesh. Several rice types are the primary focus of this study. Aus, Aman, and Boro are their names. Yield Predictions for Aus, Aman, and Boro Rice via Data Mining and ML. Six different regression methods were used to forecast the harvest of these plants. Ridge regression, linear regression, random forest regression, boosted regression, decision tree regression, and neural network regression are only some of the regression methods we've tested. In addition, seven different algorithms were evaluated for their ability to estimate Rice yield, and Random Forest Regression emerged as the clear victor. Our findings will pave the way for more precise estimates of future Bangladesh rice, wheat, and potato harvests.