S4VM

  S4VM 对传统的S3VM 进行了改进。传统的S3VM 基于低密度假设，它试图找到一个低密度的分界线， 也就是更倾向于决策边界穿过特征空间的低密度区域。S4VM 和S3VM 的不同点在于，S3VM 试图把注意力 集中在一个最优的低密度分界线上，而S4VM 则同时关注多个可能的低密度分界线。之所以这样做，是因为 给定一些有标记的点和大量未标记的点，可能存在着不止一个“间隔”较大的低密度分界线（如图2 ），基于 有限的标记样本，很难决定哪个是最好的。虽然这些低密度分界线都与有限的标记样本吻合，但它们之间差 异很大，因此如果选择错了，会有一个很大的损失，导致性能下降。所以S4VM 试图考虑所有可能的低密度 分界线。 (Over the past decade, many improved versions are proposed to solve every different specific problem of semi-supervised support vector machines, for example: to the inefficiency issue, meanS3VM is introduced to performance degenerate problem, S4VM is proposed)

S4VM
....\example.m,409,2011-05-09
....\libsvm-mat-2.89-3-weight
....\........................\aa.mat,4069,2010-11-09
....\........................\htm" target=_blank>COPYRIGHT,1497,2009-02-17
....\........................\example.m,286,2010-11-10
....\........................\heart_scale.mat,28904,2005-03-22
....\........................\libsvmread.c,3823,2009-04-14
....\........................\libsvmread.mexw32,8704,2010-11-09
....\........................\libsvmread.mexw64,11264,2010-11-10
....\........................\libsvmwrite.c,2123,2009-04-30
....\........................\libsvmwrite.mexw32,8192,2010-11-09
....\........................\libsvmwrite.mexw64,10240,2010-11-10
....\........................\main.m,849,2010-11-10
....\........................\make.m,743,2010-11-10
....\........................\Makefile,1462,2009-04-14
....\........................\htm" target=_blank>README,9181,2009-04-24
....\........................\svm.cpp,67704,2010-11-09
....\........................\svm.cpp.bak,67692,2010-11-09
....\........................\svm.h,3004,2010-11-09
....\........................\svm.h.bak,2986,2010-11-09
....\........................\svm.obj,164280,2010-11-10
....\........................\svmpredict.c,9050,2009-04-14
....\........................\svmpredict.mexw32,20992,2010-11-09
....\........................\svmpredict.mexw64,24576,2010-11-10
....\........................\svmtrain.c,12212,2010-11-09
....\........................\svmtrain.c.bak,12208,2010-11-09
....\........................\svmtrain.mexw32,53760,2010-11-09
....\........................\svmtrain.mexw64,68096,2010-11-10
....\........................\svm_model_matlab.c,7820,2010-11-09
....\........................\svm_model_matlab.c.bak,7810,2010-11-09
....\........................\svm_model_matlab.h,201,2007-11-23
....\........................\svm_model_matlab.obj,8905,2010-11-10
....\readme.txt,1915,2011-05-09
....\S4VM.m,9842,2011-05-09
....\wdbc10.mat,60528,2011-05-03