Hands-On Recommendation Systems with Python

Cosine similarity

In the previous chapter, we mathematically defined the cosine similarity score and used it extensively while building our content-based recommenders:

Mathematically, the Cosine similarity is defined as follows:

The cosine similarity score computes the cosine of the angle between two vectors in an n-dimensional space. When the cosine score is 1 (or angle is 0), the vectors are exactly similar. On the other hand, a cosine score of -1 (or angle 180 degrees) denotes that the two vectors are exactly dissimilar to each other. ...