Machine learning for time series forecasting : application in water surface analysis of Mun River

Abstract

At present, the world is facing with a global warming problem. Thailand is another country that has suffered from global warming as well. One of the problems that are often seen change that is of the season's Summer temperature is higher than it used to be. The rainy season does not fall according to the season. Or when the rain became a long time, it was flooded from the drainage in time at this point, if we can predict whether the amount of water to be found this year is high or low. Therefore, the study through satellite imagery was chosen as the starting point. This study methodology is consisting of data collection, data processing 1, data processing 2, and prediction. Using the principle that if the amount of water increases and the river does not look deep, the area of the river in the satellite image will also have to increase. Therefore, the study was to create a tool that can predict the river's surface area from satellite imagery. The results obtained from this study revealed that the river's surface area could indicate an increase in water content and the forecasts yield satisfactory results.