Water Level Prediction Model Using Back Propagation Neural Network Case study : The Lower of Chao Phraya Basin

Global warming is the cause of climate change effected to the severe flood disaster. Improvements of water level prediction model are needed. The accuracy of prediction model can reduce flood damage. This research aims to extend the water level prediction model with back propagation neural network. The proposed model tested the important factors in order to predict the water levels. The input of the model composes of water level, the capacity of water discharge, average rainfall runoff, height of basin at gauging station, and the maximum capacity of water discharge at gauging station. Mean Square Error and Relative Absolute Error were used for measure the accuracy of the prediction model between the actual water level and the predicted water level. The result of the prediction model has high accuracy when comparing with the actual values.