Soil Nutrient Prediction and Crop Recommendation System

Food is the necessity of humans which is in great demand. Agriculture is the major domain and primary source of food. Soil is the main resource for agriculture where the essential nutrients required for plant growth is available. Agriculture has progressed over time with modern technologies and more equipment's have replaced ancient farming methods. There is a great need to manage the amount of fertilizer to be added for every crop. Without soil testing, it is very difficult to ensure the right application of fertilizers for the crop and get the optimum yield. Analyzing soil samples in a laboratory is a time-consuming process that entails numerous steps, such as gathering soil samples, preparing them for analysis, conducting chemical or physical analysis, interpreting the analysis results, and ultimately providing fertilizer recommendations for crops. In this work, soil nutrient prediction system is developed which predicts the Nitrogen, Phosphorus, Potassium (NPK) values and recommends the crop to be cultivated. In this system, BORUTA regression model gives highest accuracy of 91%. Python is used to develop this prediction and recommendation system.