SDLMS Least mean square classifier (regression on labels)P=SDLMS(DATA) P=SDLMS(DATA,options) INPUT DATA training dataset OUTPUT P linear discriminant DESCRIPTION SDLINEAR implements linear discriminant performing least mean square regression of class labels. SDLMS can be used to build regression model for arbitrary numerical targets with 'targets' option. If sample weights are available, 'weights' option may be used to use them in projection estimation. For classification, two or more classes are required. For regression, class information is not used. OPTIONS 'target',T Regress targets in numerical matrix T (rows are samples) 'costs',C Provide a vector of class costs (if regressing labels) 'weights',W Vector W with optional per-sample weights SEE ALSO SDFISHER, SDNMEAN, SDGAUSS