2021 Volume 77 Issue 2 Pages I_1237-I_1242
It has been reported that machine learning predicted high-water levels of flood events. In this study, several approaches of water level prediction were carried out based on machine learning models for the Chikuma River suffering huge damage in the typhoon No. 19 of 2019. Deep learning and linear regression models indicated the high accuracy of the prediction of extensive high-water levels under the limited condition. Furthermore, it was suggested that AI predicted the incremental water levels based on the water level data of observation stations, and is able to predict flood waves at high accuracy, complementary applying the flood peaks and the starting time of water level rising acquired from accumulated rainfall into the prediction.