[1]何慧,胡小红,覃华,等.用核K-means聚类减样法优化半定规划支持向量机[J].江西师范大学学报(自然科学版),2013,(06):574-578.
　HE Hui,HU Xiao-hong,QIN Hua,et al.Using Kernel K-Means Clustering Reducing Method for the Optimization of Semi-Definite Programming SVM[J].Journal of Jiangxi Normal University:Natural Science Edition,2013,(06):574-578.
点击复制

用核K-means聚类减样法优化半定规划支持向量机()
分享到：

《江西师范大学学报》（自然科学版）[ISSN:1006-6977/CN:61-1281/TN]

卷:
期数:
2013年06期
页码:
574-578
栏目:
出版日期:
2013-12-31

文章信息/Info

Title:
Using Kernel K-Means Clustering Reducing Method for the Optimization of Semi-Definite Programming SVM
作者:
何慧;胡小红;覃华;张敏
江西师范大学商学院电子商务系,江西南昌,330022；抚州市党校,江西抚州,344000；广西大学计算机与电子信息学院,广西南宁,530004
Author(s):
HE Hui；HU Xiao-hong；QIN Hua；ZHANG Min
关键词:
支持向量机; 半定规划; 核K-means聚类; 减样
Keywords:
SVM; semidefinite programming; Kernel K-means clustering; reducing
分类号:
TP309
文献标志码:
A
摘要:
提出了使用核空间K-means聚类算法从训练集中抽取特征边界支持向量集,在边界集上构造支持向量机的半定规划问题,由于边界集的规模比原始训练集要小,降低了半定规划支持向量机的规模,达到优化向量机的目的.在UCI数据集上的实验结果表明:所提优化方法在求解多核半定规划向量机时,比原始方法获得几倍以上的速度提升,分类精度基本不变.
Abstract:
Kernel K-means clustering method is proposed for abstracting the border support vector data set from training data set.The semi-definite programming SVM is solved on border set.The SVM scale is reduced as the border set is less than the original training data set,and the optimization of semi-definite programming is implemented.The experimental results on UCI data set show that the new SVM training time is several times less than the original one and the classification accuracy of new SVM is equals to original one.

参考文献/References:

[1] Bao Yukun,Liu Zhitao.A fast grid search method in support vector regression forecasting time series [C].Berlin:Springer Berlin Heidelberg,2006:504-511.
[2] Lazar A.Income prediction via support vector machine [EB/OL].
[2012-10-18].http://dblp.uni-trier.de/rec/bibtex/conf/icmla/Lazar04.
[3] Chih-Jen Lin.Binary-class cross validation with different criteria [EB/OL].
[2010-02-04.http://www.csie.ntu.edu.tw/～cjlin/libsvmtools/eval/index.html.
[4] 郑成勇.小波与多小波变换在支持向量机人脸识别中的研究 [J].江西师范大学学报:自然科学版,2008,32(1):77-80.
[5] Gert Lanckriet,Nello Cristianini,Peter Bartlett,et al.Learning the Kernel matrix with semidefinite programming[J].Journal of Machine Learnig Research,2004(5):27-72.
[6] Li Kan.A new algorithm for solving multiple Kernel problem as SLIP [C].Proceedings of the 2008 3rd International Conference on Innovative Computing Information and Control,2008:409-415.
[7] Downs T,Gates T E,Masters A.Exact simplification of support vector solutions [J].Journal of Machine Learning Research,2001(2):293-297.
[8] Haasdonk B.Feature space interpretation of SVMs with indefinite kernels [J].Pattern Analysis and Machine Intelligence,IEEE Transactions,2005,27(4):482-492.
[9] Jair Cervantes,Li Xiaoou,Yu Wen,et al.Support vector machine classification for large data sets via minimum enclosing ball clustering [J].Neurocomputing,2008(416),71:611-619.
[10] Nello Cristianini,Jaz Kandola,Andre Elisseeff,et al.On Kernel target alignment [J].Innovations in Machine Learning,2006,194:205-256.
[11] Yang Shuzhong,Luo Siwei.Learning SVM Kernel with semi-definite programming [J].Advances in Natural Computation,2005,3601:710-715.
[12] Makoto Yamashita,Katsuki Fujisawa,Masakazu Kojima.Implementation and evaluation of SDPA 6.0(semi-definite programming algorithm 6.0)[J].Optimization Methods and Software,2003,18(4):491-505.

相似文献/References:

[1]滕少华,胡俊,张巍,等.支持向量机与哈夫曼树实现多分类的研究[J].江西师范大学学报(自然科学版),2014,(04):383.
　TENG Shao-hua,HU Jun,ZHANG Wei,et al.The Research of Multi-Classification Based on SVM and Huffnan Tree[J].Journal of Jiangxi Normal University:Natural Science Edition,2014,(06):383.

备注/Memo

备注/Memo:
国家自然科学基金(61063032)；教育部人文社会科学研究规划基金(1YJAZH080)
更新日期/Last Update: 1900-01-01