Prediction of flooding in the downstream of the Three Gorges Reservoir based on a back propagation neural network optimized using the AdaBoost algorithm

Flooding is a natural disaster that threatens people’s lives and causes economic losses. The accurate prediction of water level is of great significance for flood prevention. This study aimed to predict water levels in Wuhan City, which is located in the downstream of the Three Gorges Reservoir Region. In order to improve the accuracy of flood prediction, the AdaBoost algorithm was used to optimize a traditional back propagation neural network (BPNN) in order to resolve the slow convergence speed and local minimum in water level prediction. The improved BPNN was then employed to predict the water level in the study area for prediction intervals of 1 h, 3 h, and 5 h, respectively. Compared with the original BPNN, a generalized regression neural network, and a combination of a genetic algorithm and the original BPNN, the improved BPNN achieved superior water-level prediction. Additionally, the performance of the constructed model was evaluated using the mean absolute error, root-mean-square error (RMSE), mean absolute percentage error (MAPE), the correlation coefficients between the predicted and actual values of water level, and the frequency histograms of the prediction error. The results indicate that the improved BPNN model had a lower prediction error and show a reasonable normal distribution. Therefore, it is concluded that this model is suitable for the prediction of water level.