PolyU IR

PolyU Institutional Repository >
Civil and Environmental Engineering >
CEE Journal/Magazine Articles >

Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/2315

Title: Data-driven models for monthly streamflow time series prediction
Authors: Wu, C. L.
Chau, Kwok-wing
Subjects: Hydrologic time series
Auto-regressive moving average
Artificial neural networks
Phase space reconstruction
False nearest neighbors
Dynamics of chaos
Issue Date: Dec-2010
Publisher: Elsevier Ltd.
Citation: Engineering applications of artificial intelligence, Dec. 2010. v. 23, no. 8, p. 1350-1367.
Abstract: Data-driven techniques such as Auto-Regressive Moving Average (ARMA), K-Nearest-Neighbors (KNN), and Artificial Neural Networks (ANN), are widely applied to hydrologic time series prediction. This paper investigates different data-driven models to determine the optimal approach of predicting monthly streamflow time series. Four sets of data from different locations of People’s Republic of China (Xiangjiaba, Cuntan, Manwan, and Danjiangkou) are applied for the investigation process. Correlation integral and False Nearest Neighbors (FNN) are first employed for Phase Space Reconstruction (PSR). Four models, ARMA, ANN, KNN, and Phase Space Reconstruction-based Artificial Neural Networks (ANN-PSR) are then compared by one-month-ahead forecast using Cuntan and Danjiangkou data. The KNN model performs the best among the four models, but only exhibits weak superiority to ARMA. Further analysis demonstrates that a low correlation between model inputs and outputs could be the main reason to restrict the power of ANN. A Moving Average Artificial Neural Networks (MA-ANN), using the moving average of streamflow series as inputs, is also proposed in this study. The results show that the MA-ANN has a significant improvement on the forecast accuracy compared with the original four models. This is mainly due to the improvement of correlation between inputs and outputs depending on the moving average operation. The optimal memory lengths of the moving average were three and six for Cuntan and Danjiangkou, respectively, when the optimal model inputs are recognized as the previous twelve months.
Description: DOI: 10.1016/j.engappai.2010.04.003
Rights: Engineering Applications of Artificial Intelligence © 2010 Elsevier Ltd. The journal web site is located at http://www.sciencedirect.com.
Type: Journal/Magazine Article
URI: http://hdl.handle.net/10397/2315
ISSN: 0952-1976
Appears in Collections:CEE Journal/Magazine Articles

Files in This Item:

File Description SizeFormat
EAAI6.pdfPre-published version537.42 kBAdobe PDFView/Open
Locate publisher version via PolyU eLinks

Facebook Facebook del.icio.us del.icio.us LinkedIn LinkedIn

All items in the PolyU Institutional Repository are protected by copyright, with all rights reserved, unless otherwise indicated.
No item in the PolyU IR may be reproduced for commercial or resale purposes.


© Pao Yue-kong Library, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong
Powered by DSpace (Version 1.5.2)  © MIT and HP
Feedback | Privacy Policy Statement | Copyright & Restrictions - Feedback