A Comparative Study on Machine Learning Models for a Crop Recommendation System

Agriculture contributes significantly in shaping the GDP of a nation. Farming is a crucial aspect in agricultural nations such as that of India. While numerous technological innovations have emerged in agriculture, the current need of the hour is the continual enhancement and improvement of the current technology at hand. Due to the diverse climatic conditions, an efficient technique becomes crucial for enabling better crop cultivation and supporting farmers in their production and management. Offering a recommendation system to farmers based on historic data can significantly assist in their crop cultivation. In this paper, analysis of crop dataset and crop recommendation is done using a variety of Machine Learning algorithms. The models used include traditional models such as Decision Tree Classifier, k-Nearest Neighbor and Naïve-Bayes Classifier. Ensemble Learning models such as Random Forest Classifier, a deep learning model, Multi-layer Perceptron Classifier as well as a hybrid model which combines the Decision Tree Classifier, Random Forest Classifier, Naïve-Bayes Classifier, Multi-Layer Perceptron Classifier and a Support Vector Classifier are also utilized in this paper. The models are evaluated, and their performance are assessed based on various evaluation metrics. The evaluation metrics include accuracy, precision, and recall, and F1 score of the models.