Machine Learning for Price Prediction for Agricultural Products

Abstract

Family farms play a role in economic development. Limited in terms of land, water and capital resources, family farming is essentially characterized by its use of family labour. Family farms must choose which agricultural products to produce; however, they do not have the necessary tools for optimizing their decisions. Knowing which products will have the best prices at harvest is important to farmers. At this point, machine learning technology has been used to solve classification and prediction problems, such as price prediction. This work aims to review the literature in this area related to price prediction for agricultural products and seeks to identify the research paradigms employed, the type of research used, the most commonly used algorithms and techniques for evaluation, and the agricultural products used in these predictions. The results show that the mostly commonly used research paradigm is positivism, the research is quantitative and longitudinal in nature and neural networks are the most commonly used algorithms.