Adaptive neuro fuzzy inference system based multicrop yield prediction in the semi arid region of India

India, with a population of 1.43 billion, is the most populous country in the world, necessitating more significant food production. To ensure this food production, the country's farmers must focus on high-yielding and draught-tolerant varieties of the crops. However, just keeping an eye on this will not solve the purpose. Reliable crop yield estimation well before the sowing season is essential for planning and management in light of the changing climate. Knowing the expected yield of one's standing crops is crucial to farmers and can be a complicated task in and of itself. In this regard, modern artificial intelligence algorithms have shown to be handy tools for accurately predicting agricultural production. With this view, in this present study and attempt was made to develop multi-crop yield prediction models using an Adaptive Neuro Fuzzy Interference System (ANFIS). The crop yield is significantly influenced by climatic variables such as rainfall, minimum and maximum temperatures, relative humidity, and evaporation. Therefore, these variables were selected as input parameters, and yield of five significant crops, Kharif rice, sorghum, maize, groundnut, and Sugarcane, were selected as the output for the model development. With acceptable accuracy, the developed models have functioned well. The association between climatic variables and the agricultural production of the crops under study was disclosed by the ANFIS, and the accuracy of the rule validated this relationship. It was found that the rules formed in ANFIS accurately predicted each crop's crop yield and, therefore, emerged as a novel technique to predict seasonal multi-crop yield. It is also proposed that the ANFIS technique is best suited to predict the seasonal multi-crop yield of the semi-arid region of India.