THE RELATIONSHIP MODELS BETWEEN THE EXTREME VALUES OF RAINFALL AND TEMPERATURE WITH THE YIELD OF SAFE RICE IN PHITSANULOK PROVINCE

Abstract

         The ability to predict the future crop yield facilitates the responsible authorities to make the most appropriate decisions in order to ensure food security in the present and the future. As a result, studying the relationship between climate extremes and agricultural productivity is essential. Therefore, this research was to study the relationship between the extreme values of rainfall and temperature and the yield of safe rice in Phitsanulok province using three machine learning methods. Three machine learning methods consisted of multiple linear regression, random forest and support vector machine.  The results showed that the extreme values of rainfall correlate with safe rice yields in Phitsanulok province. The extreme values of rainfall could explain 86% of the variation in the yield of safe rice in Phitsanulok province. The random forest method was an effective and reliable method for this modeling. In addition, it was found that the extreme values of average temperature were related to the yield of safe rice in Phitsanulok province. The extreme values of average temperature could explain the variation in safety rice yield in Phitsanulok province by 98%. The support vector machine method was effective and reliable for modeling the yield of safe rice in Phitsanulok province.