标题:
基于PPR和SVM模型研究洞庭湖径流与输沙量变化Study on Runoff and Sediment Variation of the Dongting Lake Based on PPR and SVM Model
作者:
赵姗, 周念清, 李正最
关键字:
洞庭湖, 径流量, 输沙量, 投影寻踪回归(PPR), 支持向量机(SVM)Dongting Lake; Runoff; Sediment Variations; Projection Pursuit Regression; Support Vector Machine
期刊名称:
《Journal of Water Resources Research》, Vol.1 No.5, 2012-10-31
摘要:
洞庭湖是长江中游典型的吞吐型调蓄湖泊。受自然和人为因素影响，其输入和输出的径流与输沙量关系先后发生了多次大的调整过程。为了探明洞庭湖水沙出入湖量变化和相互关系，本文利用现有水文泥沙等观测试验资料，基于投影寻踪回归(PPR)和支持向量机模型(SVM)对洞庭湖径流与输沙量变化特征进行了模拟和验证，并对模拟误差进行了对比。结果表明，两种模型均可以用于模拟洞庭湖水网径流与输沙量关系，但支持向量机模型(SVM)精确度较高，适用性较好，可为洞庭湖的综合整治提供理论支撑和科学依据。
The Dongting Lake is very important for flood storage and water sources in the midstream of the Yangtze River. Due to the dual effects of nature and human activities, several major changes have been taken place on the runoff and sediment conditions and the relationships between rivers and lakes successively. In order to explore the relationships of runoff and sediment in and out the lake, the existing hydrologic and sediment and other observation data of the Dongting Lake area are fully used, and two non-linear simulation models, Projection Pursuit Regression (PPR) and Support Vector Machine (SVM) are established. The simulation errors are also compared. The results show that the Support Vector Machine (SVM) has better validity and credibility, which could be used as an effective tool to simulate the complicated river system. This finding provides theoretical support and scientific basis for comprehensive improvement and ecological restoration of the Dongting Lake.