Exploring the suitability of machine learning algorithms for crop yield forecasting using weather variables

Abstract

Crop yield forecast is valuable to many players in the agri-food chain, including agronomists, farmers, policymakers and merchants of commodities. Machine learning may be used to estimate crop yields, as well as to decide what crops to sow and what to do during the growing season. In present study Machine learning techniques such as Random Forest Regression and Support Vector Regression has been applied on three different datasets. Statistical indicators like Root Mean Square Error (RMSE), Mean Absolute Prediction Error (MAPE), and Mean Absolute Deviation (MAD) were used to compare the suggested models’ forecasting performance. Also comparison has been done of both the machine learning techniques with the stepwise regression method. Support Vector regression was observed as the best machine learning technique. However, performance of the popular statistical approach (Stepwise regression) was found to be in between the two-machine learning algorithm.