Forecasting yield by integrating agrarian factors and machine learning models: A survey

Abstract

The advancement in science and technology has led to a substantial amount of data from various fields of agriculture to be incremented in the public domain. Hence a desideratum arises from the investigation of the available data and integrating them with a process like a crop improvement, yield prediction, crop disease analysis, identifying water stress, and so on. Computing techniques like Machine learning is a new advent for the analysis and resoluteness of these intricate issues. Various analytical models like Decision Trees, Random Forests, Support Vector Machines, Bayesian Networks, and Artificial Neural Networks, and so on, have been utilized for engendering the models and analyze the results. These methods enable to analyze soil, climate, and water regime which are significantly involved in crop growth and precision farming. This survey incorporates an overview of some of the existing supervised and unsupervised machine learning models associated with the crop yield in literature. Moreover, this survey compares one approach with other using various error measures like Root Mean Square Error (RMSE), Relative Root Mean Square Error (RRMSE), Mean Absolute Error (MAE), and Coefficient of Determination (R2).