基于聚类和SVM的数据分类方法与实验研究  

New Method of Data Classification with Experimental Research Based on Clustered Group SVM

作　　者：梁修荣 杨正益[2] LIANG Xiu-rong;YANG Zheng-yi(Department of Information Engineering,Urban Vocational College of Chongqing,Yongchuan Chongqing 402160,China;School of Soft Engineering,Chongqing University,Chongqing 401331,China)

机构地区：[1]重庆城市职业学院信息工程系,重庆永川402160 [2]重庆大学软件学院,重庆401331

出　　处：《西南师范大学学报（自然科学版）》2018年第3期91-96,共6页Journal of Southwest China Normal University(Natural Science Edition)

基　　金：重庆市教委科学技术研究项目(KJ1603701)

摘　　要：提出一种新的数据分类方法,K-均值聚类分组支持向量机(CG-SVM)方法,在不删减任何样本信息的情况下,该方法使用轮廓系数作为类分离度量参数对数据的正、负类样本进行聚类、分组,对小组进行交叉结合训练,得到不同的SVM分类器,根据测试样本到聚类中心欧氏距离的加权来选择合适的SVM分类器进行类别判断,该方法实现了一个二分类问题转化为多个细分的二分类过程,效果由差变好.对所提方法分别进行了仿真数据和实际数据分类实验.实验结果表明,与传统的SVM方法相比,该算法可以有效地提高分类的准确率.A new data classification method,K-means clustering group support vector machine(CG-SVM)method is proposed.Under no circumstances deletion of any sample information,this method uses the silhouette coefficient as the class separation metric parameter clustering and grouping positive and negative samples.And different SVM classifiers are obtained after that the group is crossed and trained.The SVM classifier is selected according to the weight of the Euclidean distance of the test sample to the clustering center.The method achieves a two classification problem into more simple binary classification problems.And the effect is improved.The simulation data and the actual data classification experiment are carried out respectively.Compared with the traditional SVM method,experimental results show that CG-SVM method can effectively improve the accuracy of classification.

关 键 词：聚类分组SVM 轮廓系数 加权欧氏距离 实际数据分类 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]