STATSSVC Stats Support Vector Classifier (Matlab Stats Toolbox)
W = STATSVC(A,KERNEL,C,OPTTYPE)
DescriptionThis is the PRTools interface to the support vector classifier SVMTRAIN in Matlab's Stats toolbox. Like PRTools SVC it is a twoclass discriminant that also can be used for multiclass problems by an internal call to MCLASSC. SVMTRAIN uses the Sequential Minimal Optimization (SMO) method and thereby implements the L1 softmargin SVM classifier. See SVMTRAIN for more details. Nonlinear kernels have to be supplied by kernel procedures like PROXM. It is assumed that V = A*KERNEL generates the trained kernel and B*V the kernel matrix with size(B,1) rows and size(A,1) columns. Forthe radial basis kernel use PKSTATSSVC and RBSTATSSVC Like for all other PRTools classifiers, a new dataset B can be classified by D = B*W. The classifier performance can be measured by D*TESTC and the resulting labels by D*LABELD. D is a dataset with for every class a column. The values can be considered as class confidences or classifier condaitional posteriors. STATSSVM is basically a twoclass classifier. Multiclass problems are internally solved using MCLASSC resulting in a base classifier per class. The final result may be improved significantly by using a nonlinear trained combiner, e.g. by calling W = A*(STATSSVM*QDC([],[],1e6); Alternative SVM classifiers in PRTools are based on SVC and LIBSVC. See alsodatasets, mappings, svmtrain, svc, libsvc, mclassc, pkstatssvc, rbstatssvc, qdc, testc, labeld,
