comparison toolboxes/FullBNT-1.0.7/KPMstats/logist2FitRegularized.m @ 0:e9a9cd732c1e tip

first hg version after svn

+function [net, niter] = logist2FitRegularized(labels, features, maxIter)
+
+if nargin < 3, maxIter = 100; end
+
+[D  N] = size(features);
+weightPrior = 0.5;
+net = glm(D, 1, 'logistic', weightPrior);
+options = foptions;
+options(14) = maxIter;
+[net, options] = glmtrain(net, options, features', labels(:));
+niter = options(14);
+%w = logist2Fit(labelsPatches(jValidPatches), features(:, jValidPatches));
+