Mean Reciprocal Rank 

The next evaluation metric we are going to look at is MRR. MRR can be used where a model produces a list of outcomes and it measures the overall quality of the rankings. Let's first look at the equation:

As you can see, it is an average of the sum of the inverse of the ranks. Consider the following example:

In the first example, the correct genre was the second in rank, so the reciprocal rank is 1/2.  The second example's correct genre was the first in rank, so the reciprocal rank is 1/1, which is 1. Following this process, ...

Get C# Machine Learning Projects now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.