A Machine Learning and Web-Based Autonomous Agricultural System for Crop and Fertilizer Recommendation and Yield Prediction

Abstract

In the world, agriculture is recognized as a crucial economic factor since it meets one of civilization's most basic necessities, namely food. In the majority of nations, it is recognized as the primary source of revenue. Bangladesh is one of many countries that still use agricultural practices. In FY 2020–21, agricultural products accounted for about 12.5% of Bangladesh's GDP, while 40% of the population worked in agriculture. Therefore, the lack of acreage and the depletion of natural resources make it even more urgent to achieve optimal precision agriculture technologies, such as crop selection, fertilizer suggestion, water management, crop yield prediction, etc. This web-based solution uses machine learning to forecast production while also recommending crops and fertilizers. The accuracy of the machine learning models is remarkable. The maximum accuracy in the dataset for both crop recommendation and yield prediction are 99%, and it is 98% for fertilizer recommendations. We developed a farmer-friendly system with a graphical user interface to help farmers solve agricultural difficulties. This technology is very crucial to the country and can improve the traditional agricultural system. Along with using this strategy, the farmer will be able to benefit even further from farming.