I'm evaluating a few retrieval models using the 11-point Interpolated precision x recall curve (here is a description of the protocol I've been using: https://nlp.stanford.edu/IR-book/html/htmledition/evaluation-of-ranked-retrieval-results-1.html) evaluated for the Top-k. The problem is, some queries have no relevants within the Top-k. As such, I'm having issues computing the recall, and therefore the 11-point interpolated curve.

I'm leaning towards setting all values of interpolated precision as 0.0, with exception of Recall = 0.0, which should have an interpolated precision of 1.0. Is this the correct way to do it?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.