A Hybrid Approach for Crop Yield Prediction using Supervised Machine Learning

Abstract

Agricultural production has always been a vital factor in economic development, and it has had an enormous impact on our economic prosperity. Also, as science progresses rapidly, the farming industry has become one of the most critical segments to face issues in farming, such as land, groundwater flow, catastrophic events, herbicides, and pesticides. Throughout all stages of yield development, the amount of acceptable precipitation observed is critical to the development of harvests in farming. There will be occasions when the excellent monsoon season is insufficient to aid agricultural production, and understanding it can assist farmers in determining the volume of moisture that can only be made available via irrigated agriculture. Forecasting feasible precipitation and harvest and water requirements is a difficult task requiring a thorough and reliable scrutiny of a long set of variables, such as relative humidity and temperature. In the earlier period, the viable monsoon was calculated by factoring in 3 significant aspects: moisture, heat, and rain. Analyze several research mechanisms throughout history, and investigate a considerable proportion in heavy rains. We employ a hybrid approach that combines logistic regression and random forest (LRRF) to anticipate crop production concerning annual rainfall as in the research design.