0
$\begingroup$

Given this simple data:

enter image description here

How can I plot an ROC curve in Microsoft Excel? Step by step instructions would be very helpful. I know my TPR is ~87% and my FPR is ~13%. How do I know where to set the intervals on my data though?

I've already looked at this question but I'm not sure how to get the "score" of the data values.

$\endgroup$

closed as off-topic by Michael Chernick, kjetil b halvorsen, jbowman, gung Mar 6 '18 at 17:16

This question appears to be off-topic. The users who voted to close gave this specific reason:

  • "This question appears to be off-topic because EITHER it is not about statistics, machine learning, data analysis, data mining, or data visualization, OR it focuses on programming, debugging, or performing routine operations within a statistical computing platform. If the latter, you could try the support links we maintain." – Michael Chernick, kjetil b halvorsen, jbowman, gung
If this question can be reworded to fit the rules in the help center, please edit the question.

  • $\begingroup$ You don't have enough information to plot an ROC curve, in Excel or anything else. If this is all the infomration you have, then your task is impossible. $\endgroup$ – Peter Flom Mar 5 '18 at 12:45
  • $\begingroup$ @PeterFlom How would I go about finding the predicted probabilities for each data point? $\endgroup$ – fi12 Mar 5 '18 at 15:09
3
$\begingroup$

With the data you given, it is not possible to plot ROC curve.

This is because ROC curve needs predicted probabilities for each data point, so one can change the threshold to have different predicted classes. For each threshold, there re TPR and FPR associated with that, these are points in the ROC curve.

For your case, you do not have predicted probability for each data point, but have predicted label for one selected threshold, this will only be one point in the ROC curve. And generating the whole curve is not possible.

$\endgroup$
  • $\begingroup$ Would I have to use a logistic regression to find the predicted probabilities? $\endgroup$ – fi12 Mar 5 '18 at 15:03
  • $\begingroup$ @fi12 most models can produce probabilities. not necessary to be logistic regression. $\endgroup$ – hxd1011 Mar 5 '18 at 15:12
1
$\begingroup$

To plot roc curve you need not predicted classed, but predicted probabilities. Then at each possible threshold you calculate TPR and FPR and plot them.

In your case you have only one value for TPR and FPR, this means you can plot only a single point.

$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.