Development of a decision support system for planning durian production in Uttaradit Province with forecasting techniques

Abstract

The objectives of this research were to 1) measure the performance of the forecast model of durian production, and 2) develop a decision support system for planning durian production in Uttaradit province with forecasting techniques. This research uses data on durian production in Uttaradit Province: harvesting area, average production price, and information on factors related to the amount of production, such as average rainfall, average temperature, average maximum temperature, average minimum temperature, average high wind speed, and gusts of wind during 2010-2023 for forecast modeling. A forecast model was developed using 4 techniques: Artificial Neural Network, Naïve Bayes, K-Nearest Neighbor, and Decision Tree. The performance evaluation results of the forecasting model were obtained by finding accuracy from the Mean Absolute Error (MAE), Mean Square Error (MSE) and Root Mean Squared Error (RMSE). Performance tests revealed that the Decision Tree model had the lowest Mean Absolute Error, Mean Square Error and Root Mean Squared Error. Next was the K-Nearest Neighbor model, followed by Naïve Bayes and Artificial Neural Networks, respectively. Therefore, the Decision Tree modeling technique was applied in the development of the decision support system. The results of the system satisfaction evaluation found that the system usability evaluation results were at a good level with an average score of 4.49