Hands-On Recommendation Systems with Python

Clustering

In our weighted mean-based filter, we took every user into consideration when trying to predict the final rating. In contrast, our demographic-based filters only took users that fit a certain demographic into consideration. We saw that the demographic filters performed poorly compared to the weighted mean filter.

But does this necessarily imply that we need to take all users into consideration to achieve better results?

One of the major drawbacks of the demographic filters was that they were based on the assumption that people from a certain demographic think and rate alike. However, we can safely say that this is an overreached assumption. Not all men like action movies. Nor do all children like animated movies. Similarly, it ...