进化支持向量机模型及其在水质评估中的应用

PDF下载HTML

Title:

Evolutionary support vector machine model and its application in water quality assessment

作者:

钱云^1,2, 梁艳春¹, 翟天放³, 刘洪志⁴, 时小虎¹

1. 吉林大学 计算机科学与技术学院, 吉林 长春 130012;
2. 北华大学 电气信息工程学院, 吉林 吉林 132021;
3. 吉林省水利科学研究院, 吉林 长春 130022;
4. 吉林省计算中心 吉林省计算机技术研究所, 吉林 长春 130012

Author(s):

QIAN Yun^1,2, LIANG Yanchun¹, ZHAI Tianfang³, LIU Hongzhi⁴, SHI Xiaohu¹

1. College of Computer Science and Technology, Jilin University, Changchun 130012, China;
2. College of Electrical and Information Engineering, Beihua University, Jilin 132021, China;
3. Jilin Water Resources Research Institute, Changchun 130022, China;
4. Computing Center of Jilin Province, Computer Technology Research Institute of Jilin Province, Changchun 130012, China

关键词:

水质评估模型; 支持向量机(SVM); 遗传算法(GA); 径向基核函数; 惩罚因子

Keywords:

water quality assessment model; support vector machine (SVM); genetic algorithms (GA); radial basis kernel function; penalty factor

分类号:

TP391.4

DOI:

10.11992/tis.201410018

文献标志码:

A

摘要:

水质评估模型是进行水质规划、环境水污染控制和环境管理的有效工具。利用遗传算法(GA)对支持向量机(SVM)分类算法的径向基核函数参数σ和错分惩罚因子C进行组合优化,建立进化支持向量机模型,并将该模型应用于水质评估中。将该模型分别应用于松花江松原段、松花江哈尔滨段、黄河甘肃段和吉林桦甸关门砬子水库的真实数据上进行测试。实验结果表明,提出的进化支持向量机水质评估模型在分类精度和泛化能力上较经典SVM方法都有所提高,表明了该方法的有效性。

Abstract:

A water quality assessment model is an effective tool for water quality planning, environmental water pollution control and environment management. In this paper, an evolutionary support vector machine (SVM) model is developed by using genetic algorithm (GA) to combine and optimize the radial basis kernel function parameter σ and error penalty factor C of a SVM algorithm. This model is then extended to water quality assessment. To test the effectiveness of the proposed method, it is applied to a simulation on real data of the Songyuan and Harbin sections of the Songhua River, the Gansu section of the Yellow River, and the Jilin Huadian Guanmenlizi water reservoir. Simulation results show that, compared with the classical SVM method, the classification accuracy and generalization ability of the evolutionary support vector machine model for water quality assessment are improved.