Axis of ROC Curve Labeled wrongly?

Course Q&A AI For Medicine AI for Medical Diagnosis
1

C1_W2_Lab_1_roc_curve _and_threshold

plt.xlabel(‘False Positive Rate (sensitivity)’)
plt.ylabel(‘True Positive Rate (1 - sepecificity)’)

Should it not be the other way around? FPR is (1-specificity) and TPR is sensitivity?

2

Hi @getjaidev , I think you might be right. Let me take this case to the team in charge.

Thank you!

Juan

3

Hi @getjaidev ,

My understanding is that the ROC curve is defined by “sensitivity”(y-axis) vs “1 - specificity”(x-axis), where “sensitivity” and “1 - specificity” equal “True Positive Rate(TPR)” and “False Negative Rate(FPR)”, respectively.

I would like to ask you a question based on the above.
In your assignment notebook, the x-label and y-label for the ROC curve are ‘False Positive Rate (sensitivity)’ and ‘True Positive Rate (1 - sepecificity)’ ?
I don’t think any labels were specified in the assignment, so let me first confirm the above point.

If I have misunderstood the situation, there might be misprints in the assignment.
Could you explain the situation in more detail?
It might be helpful for us.

Best regards,
Nakamura

4

Hello Nakamura… Yes. Please see lab code below…

Now let’s use matplotlib to plot the function

plt.plot(fpr,tpr)
plt.plot(0,1,‘ro’)
plt.legend([‘Estimator’])
plt.annotate(f’ ({0}, {1})',xy = (0,1))
plt.xlabel(‘False Positive Rate (sensitivity)’)
plt.ylabel(‘True Positive Rate (1 - sepecificity)’)
plt.title(‘ROC Curve for the estimator’)

5

Hi @getjaidev ,

Thank you for the explanation.
I understand the situation.

Your point was the non-graded notebook, not the assignment notebook.
I am sorry for my misunderstanding.

I also think you might be right.
Let us share this issue with the team.

Thank you.
Nakamura

6

Apologies for the confusion. Yes I meant the non-graded notebook.