I have a question about slide 79 of Week 2
With 8 relevant documents, the theoretical maximum recall@5 would be 5/8=62.5%, right?
While the theoretical maximum precision is always 100%, whatever the @ number, right?
Is that the reason why we define a “Mean Average Precision” but not a “Mean Average Recall”?
The theoretical maximum for both precision and recall is 1 i.e. 100% and this would mean zero false positives or zero false negatives. However in real world data, to get 100% is highly impractical especially considering human error.
That’s when an harmonic mean of recall and precision gives a better trade off, per se if precision score is higher, recall score will be low.
in your query here I didn’t get why you are pointing recall @5 because based on image it clearly shows its irrelevant.
