• 首页
  • 期刊简介
  • 编委会
  • 投稿指南
  • 收录情况
  • 杂志订阅
  • 联系我们
引用本文:邓伟萍,桂 超,汪 波,石 黎,关培超.基于支持向量机的水质评估模型研究[J].软件工程,2022,25(1):47-49.【点击复制】
【打印本页】   【下载PDF全文】   【查看/发表评论】  【下载PDF阅读器】  
←前一篇|后一篇→ 过刊浏览
分享到: 微信 更多
基于支持向量机的水质评估模型研究
邓伟萍,桂 超,汪 波,石 黎 ,关培超
(湖北经济学院信息与通信工程学院,湖北 武汉 430205)
dwppove@163.com; 1392979674@qq.com; 30857394@qq.com; 30941665@qq.com; 149293834@qq.com
摘 要: 针对水质评估因子的模糊性和非线性特征,且水质样本小类(如高污染水质类)因样本量少而容易导致误分的问题,深入研究了支持向量机(SVM)这一善于解决非线性问题的智能模型,设计了一种多宽度复合高斯核的支持向量机模型。该模型通过多个复合高斯核扩大和控制核函数宽度,以此扩大样本间欧氏距离与差异,以解决小类的误分问题。运用MATLAB平台对2017 年全国98 个重点断面水质周报数据进行算法对比实验,结果证实多宽度核评估模型较好地提升了SVM的分类精度,对水质分类问题是可行有效的,对其他小样本分类问题也有一定的借鉴作用。
关键词: 水质评估;多宽度高斯核; 支持向量机;参数寻优
中图分类号: TP391.4    文献标识码: A
基金项目: 湖北省教育厅科技处中青年人才项目(Q20152201);湖北省教育厅科技处指导性项目(B2016143);湖北省教育厅科技处指导性项目(B2018127).
Research on Water Quality Assessment Model based on Support Vector Machine
DENG Weiping, GUI Chao, WANG Bo, SHI Li, GUAN Peichao
(School of Information and Communication Engineering, Hubei University of Economics, Wuhan 430205, China)
dwppove@163.com; 1392979674@qq.com; 30857394@qq.com; 30941665@qq.com; 149293834@qq.com
Abstract: In view of the ambiguity and non-linear characteristics of water quality assessment factors, and the problem of small water quality samples (such as high-polluted water quality) that are easy to cause misclassification due to the small sample size, support vector machine (SVM), a smart model which is good at for solving nonlinear problems, is deeply studied. This paper proposes to design a support vector machine model with a multi-width compound Gaussian kernel. The proposed model expands and controls the width of the kernel function through multiple compound Gaussian kernels to expand the Euclidean distance and difference between samples, so that the problem of misclassification of small classes can be solved. The MATLAB platform is used to conduct algorithm comparison experiments on the water quality weekly report data of 98 key sections across China in 2017. The results prove that the multi-width kernel assessment model improves the classification accuracy of SVM, which is feasible and effective for water quality classification problems. It also provides a reference for problems of other small sample classification.
Keywords: water quality assessment; multi-width Gaussian kernel; support vector machine; parameter optimization


版权所有:软件工程杂志社
地址:辽宁省沈阳市浑南区新秀街2号 邮政编码:110179
电话:0411-84767887 传真:0411-84835089 Email:semagazine@neusoft.edu.cn
备案号:辽ICP备17007376号-1
技术支持:北京勤云科技发展有限公司

用微信扫一扫

用微信扫一扫