Improved information dissemination services for the agricultural sector in Thailand: development and evaluation of a machine learning based rice crop yield prediction system.

This article focuses on the development and evaluation of a Machine Learning (ML)-based rice yield prediction system using extensive historical datasets from Thai governmental agencies, including the Office of Agricultural Economics (OAE) and the Ministry of Agriculture and Cooperatives (MOAC). Four prediction models—Generalized Linear Model (GLM), Feed-Forward Neural Network (FFNN), Support Vector Machine (SVM), and Random Forest (RF)—were developed and assessed for accuracy, computational complexity, and usability, with the FFNN model demonstrating superior prediction accuracy and fast execution during testing despite higher training complexity. A Web-based system was created to disseminate rice yield and climate information to support decision-making by farmers, policymakers, and other stakeholders, and its usability was evaluated by 16 key users from the OAE, who found the system marginally acceptable with generally positive satisfaction but neutral feedback on efficiency and effectiveness. The study highlights the potential of ML techniques in enhancing agricultural information services in Thailand and suggests future work incorporating deep learning and broader data sources to improve crop yield predictions and smart farming practices.