Choosing the right product to recommend is very important to getting a good sale, but we must intelligently recommend, that is, check if there is a possibility that a customer may like a particular product. The collaborative filtering technique helps to solve this problem. For this we use logical concepts and mathematical techniques in the recommendation algorithms, which assist in collaborative filtering. a simple example follows:
User1 liked movie B and don't like the movie A, and User2 liked movies A and C. User3 liked movies A, B and D.
Comparing the similarity between these individuals the result of the recommendation would be:
User1 = D, User2 = D and User3 = C

I launched a package with recommendation algorithms: collaborative filtering.

source code