Application of machine learning to forecast economic crop production in agricultural supply chains: Central - Western Economic Corridor Area

Abstract

          The purpose of this research article is to apply machine learning to fo recast economic crop production in the agricultural supply chain. Conducti ng research in the Central - Western Economic Corridor, which includes Ka nchanaburi, Suphan Buri, Nakhon Pathom, and Phra Nakhon Si Ayutthaya. Consider the top 3 economic crops with the highest output in the said area: sugarcane factory, factory cassava, and first-year rice, respectively. Data of economic crop statistics for the past 10  years (2013 - 2022)  from the Office of Agricultural Economics, Ministry of Agriculture and Cooperatives. Resea rch was conducted using KNIME or Konstanz Information Miner;   an open source data analysis program. KNIME has been ranked by Gertner as a lea der in data analysis and machine learning for several years in a row.  Whe n applying machine learning to forecast the amount of economic crop pro duction in the agricultural supply chain,  there are different algorithms to predict the appropriate amount of production when considering the Mean absolute percentage error. Sugarcane prediction using the Random Forest algorithm is the most appropriate. Cassava prediction using Simple Regre ssion Tree and Gradient Boosted Trees algorithms is most suitable.  Major rice prediction using the Polynomial Regression algorithm is most approp riate. Those involved or have a stake in the agricultural supply chain of ec onomic crops in the Central - Western Economic Corridor such as entrepre neurs, agricultural support and promotion agencies in the area, can apply algorithms appropriate to economic crops to forecast quantities economic crop production.   As well as agricultural planning,  planting, distribution, and marketing stimulation to be consistent with periods of high and low p roduction of economic crops in order to create a balance of demand and s upply.