基于球结构SVM的多标签分类
|
摘要
现有多标签分类问题普遍被转换成多类分类问题,计算量较大,运行时间较长,且面对新类别加入时,拓展性较差。为此,提出一种基于球结构支持向量机的多标签分类方法。每一类别标签对应一个球域结构,提取球重叠区域的样本,依据距离差值度量样本类别相似度,确定样本所属类别。实验结果表明,该方法可以节省210 ms的训练时间,使平均查全率提高3.2%,适合大量样本分类。
Multi-label classification problem is generally converted into multi-class classification problem,and it can increase the amount of computation and extend running time.When a new class is added,the method has a poor expandability too.To solve this problem,this paper proposes a multi-label classification method based on sphere structured Support Vector Machine(SVM).Each category labels correspond a ball domain structure,this paper extracts the samples of ball overlap region,according to the distance difference measures the sample category similarity,and is to determine the category of samples.Experimental results show that the method can save 210 ms training time,improves the average recall rate of 3.2% and is suitable for a large number of sample classification.
Multi-label classification problem is generally converted into multi-class classification problem,and it can increase the amount of computation and extend running time.When a new class is added,the method has a poor expandability too.To solve this problem,this paper proposes a multi-label classification method based on sphere structured Support Vector Machine(SVM).Each category labels correspond a ball domain structure,this paper extracts the samples of ball overlap region,according to the distance difference measures the sample category similarity,and is to determine the category of samples.Experimental results show that the method can save 210 ms training time,improves the average recall rate of 3.2% and is suitable for a large number of sample classification.
引文
[1]刘瑞阳,邱卫杰.基于SVM期望间隔的多标签分类的主动学习[J].计算机科学,2011,38(4):230-232.
[2]Snoek C,Worring M,Gemert J V,et al.The ChallengeProblem for Automated Detection of 101 SemanticConcepts in Multimedia[C]//Proc.of the 14th AnnualACM International Conference on Multimedia.New York,USA:ACM Press,2006.
[3]郑伟,王朝坤,刘璋,等.一种基于随机游走模型的多标签分类算法[J].计算机学报,2010,33(8):1418-1426.
[4]Hullermeier E,Furnkranz J,Cheng Weiwei,et al.LabelRanking by Learning Pairwise Preferences[J].ArtificialIntelligence,2008,172(16):1897-1916.
[5]刘端阳,邱卫杰.基于加权SVM主动学习的多标签分类[J].计算机工程,2011,37(8):181-182,185.
[6]Tax D M J,Duin R P W.Support Vector DomainDescription[J].Pattern Recognition Letters,1999,20(11):1191-1199.
[7]朱美琳,刘向东,陈世福.用球结构的支持向量机解决多分类问题[J].南京大学学报,2003,39(2):153-158.
[8]Read J.A Pruned Problem Transformation Method forMulti-label Classification[C]//Proc.of the New ZealandComputer Science Research Student Conference.Wellington,New Zealand:[s.n.],2008.
[9]Lin Chih-Jen.LIBSVM Data:Multi-label Classifica-tion[EB/OL].(2010-11-21).http://www.csie.ntu.edu.tw/~cjlin/libsvmtools/datasets/multilabel.html.
[10]苏高利,邓芳萍.关于支持向量回归机的模型选择[J].科技通报,2006,22(2):154-158.
[2]Snoek C,Worring M,Gemert J V,et al.The ChallengeProblem for Automated Detection of 101 SemanticConcepts in Multimedia[C]//Proc.of the 14th AnnualACM International Conference on Multimedia.New York,USA:ACM Press,2006.
[3]郑伟,王朝坤,刘璋,等.一种基于随机游走模型的多标签分类算法[J].计算机学报,2010,33(8):1418-1426.
[4]Hullermeier E,Furnkranz J,Cheng Weiwei,et al.LabelRanking by Learning Pairwise Preferences[J].ArtificialIntelligence,2008,172(16):1897-1916.
[5]刘端阳,邱卫杰.基于加权SVM主动学习的多标签分类[J].计算机工程,2011,37(8):181-182,185.
[6]Tax D M J,Duin R P W.Support Vector DomainDescription[J].Pattern Recognition Letters,1999,20(11):1191-1199.
[7]朱美琳,刘向东,陈世福.用球结构的支持向量机解决多分类问题[J].南京大学学报,2003,39(2):153-158.
[8]Read J.A Pruned Problem Transformation Method forMulti-label Classification[C]//Proc.of the New ZealandComputer Science Research Student Conference.Wellington,New Zealand:[s.n.],2008.
[9]Lin Chih-Jen.LIBSVM Data:Multi-label Classifica-tion[EB/OL].(2010-11-21).http://www.csie.ntu.edu.tw/~cjlin/libsvmtools/datasets/multilabel.html.
[10]苏高利,邓芳萍.关于支持向量回归机的模型选择[J].科技通报,2006,22(2):154-158.