ssa-lssvm在中长期径流意料中的运用探究.pdfVIP

Journal of Water Resources Research 水资源研究, 2016, 5(5), 423-433 Published Online October 2016 in Hans. /journal/jwrr /10.12677/jwrr.2016.55049 Application of SSA-LSSVM in Mid-Long Term Runoff Prediction Huanhuan Ba1,2, Shenglian Guo1,2, Yixuan Zhong1,2, Zhangjun Liu1,2 1 State Key Laboratory of Water Resources and Hydropower Engineering Science, Wuhan University, Wuhan Hubei 2 Hubei Provincial Collaborative Innovation Centre for Water Resources Security, Wuhan Hubei th th th Received: Sep. 17 , 2016; accepted: Oct. 7 , 2016; published: Oct. 10 , 2016 Copyright © 2016 by authors and Hans Publishers Inc. This work is licensed under the Creative Commons Attribution International License (CC BY). /licenses/by/4.0/ Open Access Abstract To improve the accuracy of runoff prediction, Singular Spectrum Analysis (SSA) is applied to preprocess the original flow series and a new reconstructed series is obtained. The monthly inflow data of the Shui- buya Reservoir from 1951 to 2009 were selected as a case study. Seasonal Autoregressive (SAR) model, support vector machine (SVM) and least square support vector machine (LSSVM) are used to simulate and predict the original and reconstructed data series. The results show that SSA-LSSVM performs the best among these models, in which the model efficiency coefficients reach 89% and 84% during the veri- fication and testing periods, respectively. It is shown that the accuracy of mid-long term runoff predic- tion can be significantly improved by using SSA. Keywords Mid-Long Term Runoff Prediction, Singular Spectrum Analysis, SAR Model, SVM, LSSVM, Shuibuya Reservoir SSA-LSSVM在中长期径流预测中的应用研究 1,2 1,2 1,2 1,2 巴欢欢 ，郭生练 ，钟逸轩 ，刘章君 1武汉大学水资源与水电工程科学国家重点实验室，湖北 武汉 2水资源安全保障湖北省协同创新中心，湖北 武汉 收稿日期：2016年9月17 日；录用日期：2016年10月7 日；发布日期：2016年10月10 日 作者简介：巴欢欢(1992-)，女，河南灵宝人，硕士研究生，主要从事水文