Crop and Fertilizer Recommendation System Using Machine Learning

Soil and fertilizers are an integral part of the yield to be harvested since the right type of soil and fertilizers in accurate amounts can enhance both the growth and health of the crops to a substantial degree. However, the present-day system does not comprehend intelligent recommendations, but instead a scarcity (or nescience) of experimentation and legacy knowledge serve to strip farmers of additional money and time. Hence, a system that is both effective and user-friendly should be adopted. Thus, this chapter proposes and implements a system to predict suitable crops and fertilizers according to geographic location and soil quality by applying machine learning algorithms. The proposed system mainly comprises four stages: analyzing and visualizing the data; separating the data into training and testing sets; training the machine learning models; and, finally, comparing the accuracy of each model. XGBoost and Random Forest were the two models that were the most accurate when it came to determining the most efficient recommendation model. The highest accuracy was found out to be 99.31% for crops and 90% for fertilizers, making the system extremely valuable for the farmers to amplify their harvest yield.