Hi,
I want to do a classifier analysis, with 2 negative classes (classes 1 and 3). In labels, there are the true classes (1×30 double) with labels 1, 2 or 3. In scores, i measurement values (1×30 double), going from 0 to 114. I expect suby to be the result for classifying group 2 from group 1 and group 2 respectively. suby has the correct dimension of 30×2 double, and subynames is a correct 1×2 cell, with the class names 1 and 3.
But the 2 columns of suby are exactly equal and equal to y. What is wrong? I want two ROC curves of two independent classifiers, one for group 2 versus group 1 and one of group 2 versus group 3. So what is the sense behind suby?
posclass = 2; negclass = [1 3]; [x,y,t,auc,optrocpt,suby,subynames] = perfcurve(labels,scores,posclass,'negClass',negclass);
Thank you, Philipp
Best Answer