Hands-On Recommendation Systems with Python

Weighted mean

In the previous model, we assigned equal weights to all the users. However, it makes intuitive sense to give more preference to those users whose ratings are similar to the user in question than the other users whose ratings are not.

Therefore, let's alter our previous model by introducing a weight coefficient. This coefficient will be one of the similarity metrics that we computed in the previous chapter. Mathematically, it is represented as follows:

In this formula, r_u,m represents the rating given by user u to movie m.

For the sake of this exercise, we will use the cosine score as our similarity function (or sim). Recall ...