PROBLEM TO BE SOLVED: To predict a waver level of a river, a dam, sewage water, etc., with sufficiently high precision while maintaining how high and low correlation between data is when utilizing a neural network so as to predict the water level, according to the disclosure.;SOLUTION: The present disclosure relates to a water level prediction method comprising, in order: a data input step of inputting a region upstream from a water level prediction point and rainfall data before water level prediction time, and the water level prediction point and water level data after the water level prediction time; convolution steps R1-RM of executing processing in a convolution layer and a pooling layer of a convolution neural network on the rainfall data; all combined steps C1-C4 of executing processing of all combined layers of the convolution neural network on the convoluted rainfall data and water level data; and a water level prediction step of predicting water levels at the water level prediction point and water level prediction time.;SELECTED DRAWING: Figure 1;COPYRIGHT: (C)2019,JPO&INPIT
展开▼