NLP 101: Understanding Precison and Recall

If you’re dealing with NLP, you better know the terms precison and recall. If that’s not the case, however, read on! There’s always got to be a first time!

The Brain Game

Let’s say I want to test your memorization skills, so I ask you to play a brain game. The game is simple. I show you a bunch of objects, and all you have to do is remember as much of them as possible. Easy? Cool. Time’s up. Let’s test you out.

Let’s say you call out a bunch of items H = {h1, h2, … hm}, while I actually showed you T = {t1, t2, … tn}.

Recall

Recall, quite simply, is how much do you correctly recall!

So out of the H items you think you saw, you only recall H∩T of them correctly. How many of the T items did you actually remember? Yup, that’s your recall rate.

Precision

Precision is how often do you get your judgements right.

So out of the |H| attempts you make, you were right only |H∩T| times. Your precision?

Easy, right? So if you ever have a hard time with this again, just remember the brain game we played.

Cheers!

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.