Copula Entropy coupled with Wavelet Neural Network Model for Hydrological Prediction

3rd International Conference on Advances in Energy Resources and Environment Engineering

能源学;生态环境科学

Wang, Yin^1 ; Yue, Jiguang^1 ; Liu, Shuguang^2 ; Wang, Li^1

School of Electronics and Information Engineering, Tongji University, Shanghai

201804, China^1

School of Civil Engineering, Tongji University, Shanghai

201804, China^2

关键词: Dynamical pattern;    Hydrological forecasting;    Hydrological prediction;    Linear correlation;    Mutual informations;    Nonlinear fitting;    Wavelet neural network model;    Wavelet neural networks;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/113/1/012160/pdf
DOI  :  10.1088/1755-1315/113/1/012160

学科分类：环境科学（综合）

来源: IOP

【 摘 要 】

Artificial Neural network(ANN) has been widely used in hydrological forecasting. in this paper an attempt has been made to find an alternative method for hydrological prediction by combining Copula Entropy(CE) with Wavelet Neural Network(WNN), CE theory permits to calculate mutual information(MI) to select Input variables which avoids the limitations of the traditional linear correlation(LCC) analysis. Wavelet analysis can provide the exact locality of any changes in the dynamical patterns of the sequence Coupled with ANN Strong non-linear fitting ability. WNN model was able to provide a good fit with the hydrological data. finally, the hybrid model(CE+WNN) have been applied to daily water level of Taihu Lake Basin, and compared with CE ANN, LCC WNN and LCC ANN. Results showed that the hybrid model produced better results in estimating the hydrograph properties than the latter models.