Crop Prediction using Machine Learning Techniques and IoT

Agriculture being the backbone of the Indian economy faces multiple challenges in the production of quality food and good yield due to the unnatural climatic changes, which reduces crop yield adversely. Farmers lack the knowledge to forecast the future crop yield and take the necessary precautions. Therefore, this research aims to assist farmers in crop selection by predicting suitable crops for a given region using supervised machine learning algorithms. The work focusses on achieving the improved accuracy of crop prediction when compared to the existing models. Seed data with the appropriate parameters such as humidity, soil pH, temperature, rainfall, nitrogen, potassium, and phosphorus content are collected to support its maximum growth. Based on the obtained data, the crop prediction is performed. Performance of the Naïve Bayes Classifier, Ada Boost Classifier, Decision Tree Classification, and Voting Classifier algorithms are compared based on the accuracy, confusion matrix, and other performance metrics like precision, recall, and F1 score. The Decision Tree classifier algorithm is chosen as the best algorithm for crop prediction and the model is hence deployed with it. The finest accuracy of around 99.4 % is achieved by the proposed model.