-1

I have True positive rate and False positive rate for different thresholds. Now I need to compute AUC_ROC curve without using sckit-learn library.

TPR and FPR values are below:
TPR = [0.0001,0.0002,0.0003,0.0004,0.0005,0.0006,0.0007,0.0008,0.0009,0.001]
FPR = [0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0]

Can someone please tell how to compute the scores. I used numpy.trapz(TPR,FPR). But the output is weird. Can you suggest how to do this?

Rajeev Srivastava
  • 159
  • 2
  • 12

1 Answers1

0

perhaps, the input arrays should have the boundary points ( (0,0) and (1,1) ) included for np.trapz to work correctly. Also, having all FPR equal to zero seems strange

user3294756
  • 1