Crop selection using machine learning techniques for better agriculture performance

With the development of innovations like machine learning, the agricultural industry has been the focus of substantial study in the past few years. Choosing the best crop to use the area efficiently is a challenging concern. It directly impact to the producers owing of the ever-changing economics of Agri-produce. Crop production is heavily influenced by environmental elements like as the composition of the soil, temperature, humidity, rainfall, climate, or the area under cultivation, among other necessary measurements. These elements have an active or passive effect on crop yield. With multiple options of crops, it is essential for the farmers to plan the crop growing strategy in advance. If the farmer gets the advance estimate of each type of the crop yield in advance, cultivation can be done accordingly. Proposed work emphasizes to analyse the impact of machine learning algorithms to estimate the crop quantity as a final outcome. It can be used to give insights to the farmer about optimizing resource allocation and boost overall crop yield. To solve the crop selection problem, machine learning (ML) algorithms like Logistic Regression, Gaussian Naïve Bayes, SVC, K-Nearest Neighbors, Decision Tree, Extra Tree, Random Forest, Bagging Classifier, Gradient Boosting Classifier and Ada boosting classifier are implemented to get the accurate classification results. Total 2200 samples are taken for the implementation purpose. Out of 2200, 1760 samples are considered for the training and 440 samples are reserved for the testing. With an accuracy of 99.5% precision of 99.6% recall of 99.5% and the F1 score of 99.5%, the Gaussian Naive Bayes model scores most effectively out of all the models used in yield prediction