Not a member of Pastebin yet?
Sign Up,
it unlocks many cool features!
- %% PLOT RECEIVER OPERATING CHARACTERISTIC (ROC)
- % ROC is a metric of classifier quality
- simTRAIN = sim(net,TRAINMX);
- simTEST = sim(net,TESTMX);
- [TPR,FPR,THR] = roc(TRAINLAB,simTRAIN);
- % TPR : true-positive/positive ratios. (guess_case|is_case)/guess_case
- % FPR : false-positive/negative ratios. (guess_case|is_ctrl)/guess_ctrl
- % THR : thresholds over interval [0,1]
- TPR = cell2mat(TPR');
- FPR = cell2mat(FPR');
- THR = cell2mat(THR');
- close all
- fh1=figure('Units','normalized','OuterPosition',[.1 .06 .6 .9],'Color','w');
- ax1 = axes('Position',[.1 .1 .8 .8],'Color','none'); hold on
- ph1 = plot(FPR',TPR','LineWidth',5);
- ph2 = line([0 1],[0 1],'Color','k','LineStyle','--');
- Pfilter=.01; box on; grid on
- title(['ROC P(num2str(Pfilter))=' num2str(size(TRAINMX,1)) ' variants'])
Add Comment
Please, Sign In to add comment