-1

I have applied Random Forest algorithm on the dataset having 203 classes. I have applied 100 fold cross validation. The metric i used is auroc. what does this auroc graph represents that I got given below?AUROC graph of random forest with 100 fold cross validation on 203 class dataset

desertnaut
  • 57,590
  • 26
  • 140
  • 166
AkAnKsHa BaLi
  • 37
  • 1
  • 1
  • 12
  • I know that, highér the lines above median lines, tpr increases – AkAnKsHa BaLi Oct 07 '22 at 16:09
  • 1
    I am afraid interpreting ROC curves is not a *programming* question, hence it is off-topic here; please see the intro and NOTE in https://stackoverflow.com/tags/machine-learning/info – desertnaut Oct 07 '22 at 16:09
  • It is also seen from the graph that many classes overlapped with each other it means is not a good classifier because it is unable to do separation of various classes. – AkAnKsHa BaLi Oct 07 '22 at 16:14
  • 1
    Please notice that commenting like that in your own posts does not make much sense; if you want to add information, please edit & update your question accordingly. – desertnaut Oct 07 '22 at 17:14

1 Answers1

0

This AUROC graph represents the overlapping of different classes. So it is working as a bad classifier that is unable to distinguish different classes accurately. The higher the graph is above the middle linear line means higher the positive classes are classified accurately and have less specificity. The right top side graph shows high sensitivity but a high false positive rate ( less specificity). The left top side graph shows high sensitivity and high specificity (less false positive rate)

AkAnKsHa BaLi
  • 37
  • 1
  • 1
  • 12