TY - JOUR
T1 - Support vector regression for real-time flood stage forecasting
AU - Yu, Pao Shan
AU - Chen, Shien Tsung
AU - Chang, I. Fan
N1 - Funding Information:
The authors thank the National Science Council of the Republic of China, Taiwan for financially supporting this research under Contract Nos. NSC92-2625-Z-006-003 and NSC93-2625-Z-006-001.
Copyright:
Copyright 2008 Elsevier B.V., All rights reserved.
PY - 2006/9/15
Y1 - 2006/9/15
N2 - Flood forecasting is an important non-structural approach for flood mitigation. The flood stage is chosen as the variable to be forecasted because it is practically useful in flood forecasting. The support vector machine, a novel artificial intelligence-based method developed from statistical learning theory, is adopted herein to establish a real-time stage forecasting model. The lags associated with the input variables are determined by applying the hydrological concept of the time of response, and a two-step grid search method is applied to find the optimal parameters, and thus overcome the difficulties in constructing the learning machine. Two structures of models used to perform multiple-hour-ahead stage forecasts are developed. Validation results from flood events in Lan-Yang River, Taiwan, revealed that the proposed models can effectively predict the flood stage forecasts one-to-six-hours ahead. Moreover, a sensitivity analysis was conducted on the lags associated with the input variables.
AB - Flood forecasting is an important non-structural approach for flood mitigation. The flood stage is chosen as the variable to be forecasted because it is practically useful in flood forecasting. The support vector machine, a novel artificial intelligence-based method developed from statistical learning theory, is adopted herein to establish a real-time stage forecasting model. The lags associated with the input variables are determined by applying the hydrological concept of the time of response, and a two-step grid search method is applied to find the optimal parameters, and thus overcome the difficulties in constructing the learning machine. Two structures of models used to perform multiple-hour-ahead stage forecasts are developed. Validation results from flood events in Lan-Yang River, Taiwan, revealed that the proposed models can effectively predict the flood stage forecasts one-to-six-hours ahead. Moreover, a sensitivity analysis was conducted on the lags associated with the input variables.
UR - http://www.scopus.com/inward/record.url?scp=33746916489&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=33746916489&partnerID=8YFLogxK
U2 - 10.1016/j.jhydrol.2006.01.021
DO - 10.1016/j.jhydrol.2006.01.021
M3 - Article
AN - SCOPUS:33746916489
VL - 328
SP - 704
EP - 716
JO - Journal of Hydrology
JF - Journal of Hydrology
SN - 0022-1694
IS - 3-4
ER -