Weather based Localized Crop Prediction using Machine Learning

Abstract

Crop prediction in agriculture is critical and essentially depends upon soil and environmental conditions which include rainfall, humidity, and temperature. Accurate crop prediction results in increased crop production. Recently, machine leering techniques have been successfully employed in the agriculture field for classification and detection tasks. The fundamental goal of this research is to employ several machine learning methods to forecast the accurate crop for a land-based on soil and weather parameters. The classification algorithms employed in this study involve Logistic Regression, Naive Bayes, Random Forest, Support Vector Machines (SVM), XGBoost, and AdaBoost; with XGBoost offering the highest level of prediction accuracy and reliability. This work can greatly assist farmers and other stakeholders in making appropriate storage and business decisions to locate the crop before sowing. Moreover, a web-based application using Flask platform is developed to assist farmers in choosing which crop to cultivate to elicit the greatest return.