摘要
To solve the problems of low prediction accuracy, poor stability, and low calculation efficiency in runoff forecasting, this study develops an extreme learning machine (ELM) model based on improved complete ensemble empirical mode decomposition adaptive noise (CEEMDAN). The model first uses orthogonal triangular decomposition (QR) to redefine the ELM hidden layer output to construct the orthogonal triangular ELM (QR-ELM). It then introduces the online sequence mechanism (OS) into the QR-ELM model to construct online sequence OR-ELM (OS-QR-ELM), which can effectively improve the efficiency of the ELM models. Then, it uses the correlation wave extension method to extend both ends of the original signal to solve the end effect of CEEMDAN. Finally, it integrates CEEMDAN and OS-QR-ELM combined with parallel computing ideas to construct a parallel CEEMDAN-OS-QR-ELM runoff prediction method. The experimental results show that compared with the support vector regression method combined with empirical mode decomposition, the accuracy and reliability of the parallel CEEMDAN-OS-QR-ELM model was 1.09% and 8% higher, respectively. Compared with the ELM model, the efficiency of OS-QR-ELM was 5%.
学科与关键词
- Two-phase extreme learning machines integrated with the complete ensemble empirical mode decomposition with adaptive noise algorithm for multi-scale runoff prediction problemsWen, Xiaohu*; Feng, Qi; Deo, Ravinesh C.*; Wu, Min; Yin, Zhenliang; Yang, Linshan; Singh, Vijay P.Journal of Hydrology, 2019, 570: 167-184.被引 98
- Runoff prediction using a multi-scale two-phase processing hybrid modelZhao, Xuehua*; Wang, Huifang; Guo, Qiucen; An, JiatongStochastic Environmental Research and Risk Assessment, 2025.
- Forecasting reservoir monthly runoff via ensemble empirical mode decomposition and extreme learning machine optimized by an improved gravitational search algorithmNiu, Wen-jing; Feng, Zhong-kai*; Zeng, Ming; Feng, Bao-fei; Min, Yao-wu; Cheng, Chun-tian; Zhou, Jian-zhongAPPLIED SOFT COMPUTING, 2019, 82: 105589.被引 91
