Determination of Land Suitability for Rice Cultivation using Random Forest Analysis in Central Region Thailand

• Published in: 2022 3rd International Conference on Big Data Analytics and Practices (IBDAP) pp. 47 - 51
• Main Authors: Dolsiririttigul, Pattaraporn, Pichitlamken, Juta, Jaijit, Sasarose
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.09.2022
• Subjects: Big Data; clustering; Crops; Economics; Environmental factors; Forestry; land suitability; machine learning; random forest; Random forests; rice cultivation
• DOI: 10.1109/IBDAP55587.2022.9907329

Rice is one of Thailand's top five cash crops. Rice cultivation areas in Thailand's central region have gradually been changed to residential and industrial zones, resulting in lost export revenues and opportunities. Land suitability in Thailand have been performed in recent years. In this study, the random forest approach with voting ensemble supervised machine learning technique is used to identify suitable provinces in the central region of Thailand for rice cultivation. We consider economic, social, and environmental factors, and the results are integrated into a single land suitability using the agglomerative hierarchical cluster technique. We find that most of the central regions are classified as moderately suitable (40.91% or 0.76 million hectares), followed by unsuitable (31.82% or 1.67 million hectares). The areas are determined highly appropriate and marginally suitable at 22.73 percent (2.07 million hectares) and 4.55 percent (0.49 million hectares), respectively. The suitable crop areas assessment of this study can forecast future crop cultivating volumes based on the suitable area results influenced by the economic, social and environmental factors.