Implemented a recommendation system for movies using content based filtering.
In content based filtering, we suggest movies based on the content that the user prefers and then finding out similar movies using techniques like cosine distance and Euclidean distance.
We have used cosine angle method to estimate the similarity between the two documents as it has its own advantages over Euclidean distance for avoiding false positives associated with uneven documents.
Libraries used:-
- pandas
- numpy
- sklearn.feature_extraction
- sklearn.pairwise
Features considered : keywords,cast,genres,director
The provided dataset contains lots of other features as well. So, if one wants to use other features, feel free to do so.