Precipitation Prediction by Artificial Intelligence in Thung Kula Ronghai

The purposes of this research are to study and find a suitable model of Artificial Intelligence (AI) for forecasting precipitation in the area of Thung Kula Rong Hai using Single Decision Trees, Random Forests, Tree Boost and Artificial Neural Networks. The variables used in the research are temperature humidity and precipitation from Copernicus Climate Change Service website, which is daily data during January 1, 2009 to December 31, 2022 covering a period of 5113 days. Data are divided into 2 sets; the first set, Train data set (70%) for use in modeling precipitation forecasts, and the second set, Test data set (30%) for comparing the accuracy of the forecast values using Root mean square error (RMSE). It has been found that Artificial Neural Networks (ANNs) provides the best forecast of precipitation in the area of Thung Kula Rong Hai (RMSE = 0.000526). The key factor that affects precipitation is humidity.