Machine Learning / F-Beta Score

F-Beta Score

By Marcelo Fernandes Dez 7, 2017

F-Beta Score

F-Beta score is a way of measuring a certain accuracy for a model. It takes into consideration both the Recall and Precision metrics. If you don't know what those are, it's highly recommended to check the previous post about Recall & Precision.

A quick definition of recall and precision, in a non-mathematical way:

  • Precision: high precision means that an algorithm returned substantially more relevant results than irrelevant ones
  • Recall: high recall means that an algorithm returned most of the relevant results

So the F-Beta Score is defined as:

When Beta is equal to 1, we have the harmonic mean, and for this particular value of Beta, we say that this is the F1 Score:

Some Rules of Thumb:

  • To give more weight to the Precision, we pick a Beta value in the interval 0 < Beta < 1
  • To give more weight to the Recall, we pick a Beta Value in the interval 1 < Beta < +∞

In practice, you will be testing which value of Beta satisfies your solution better, and eventually, we expect you to find out the correct answer. The first step towards success is to understand your data. So what do you need? More precision, or more Recalling?