Movie Recommender web service (ML)
This is the first part of the story of building a movie recommender and deploy it. This article is only introduction to development of KNN model for movie recommender.
The repository for this article could be found here. The part 2 of this article is also here.
Data
The data-set is downloaded from MovieLens; based on these data the KNN is implemented like this in here.
This implementation is very simple, first a movie that has been selected to get similar movie to it. For this purpose, cosine similarity based on movie’s genres is computed; then popularity of selected movie to other is added to that. Then based on demanded number of similar movie to recommend; it will be sorted and to K item will return.
Now let’s deploy it with flask. In this story, the implementation of recommender web service with celery, flask, mongodb and nginx on docker is revealed.
