Automated Crop Recommender System using Pattern Classifiers

One of the biggest problems faced in the twenty-first century is how to produce food sustainably. High food productivity goals were nevertheless met by the green revolution, albeit at the expense of environmental damage such as soil deterioration and climate change. Crops are often recommended based on the weather, humidity, rainfall, and other parameters mechanically without any practical experience to increase agricultural productivity. One must have the knowledge of which crops may be planted effectively one after the other at any time since inappropriate implantation produces considerably more harm than benefit. The use of machine learning classifiers like System Vector Machine (SVM), Random Forest, Decision Tree, and KNN to forecast which crop is optimal for which soil type based on data sets (n=100) like pH and yearly rainfall values is suggested in this paper. By comparing the results obtained through these pattern classifiers and determining their total accuracy rates, the crops such as wheat and ragi can be recommended. The results show KNN to have 98% accuracy in recommending the crop. This kind of automated system can benefit not only the farmers but also anyone interested in doing their own gardening at home since they may learn what kind of crop should be grown in soil with input parameters like pH level and previous year rainfall details.