My question was closed as a duplicate of another question.

I was asking for a formula, if any, to compute accuracy for a logistic regression without using a separate test set. For instance (I am just making it up): $ \text{accuracy} = (\hat{\sigma}^2/N) \hat{y}^2 \text{Var}(\hat{\beta})$ or whatever.

The "duplicate" question would have me compute the accuracy using the bootstrap. Which of course is just averaging many test sets.

How is that an answer?

I think my question wasn't clear. Can I modify it and have the question re-opened?


Yes, you can edit your question to clarify how it differs from the proposed duplicate; in fact, that's exactly you're expected to do when the answers aren't what you wanted. In this particular case, explain that you're asking about analytic solutions, & not resampling-based estimators of out-of-sample performance.

(For what it's worth, bootstrapping is probably the best approach here, & a valid answer to the original version of the question.)

  • $\begingroup$ I have modified the question and removed the code which was misleading $\endgroup$ Apr 12 at 9:30

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .