Collebrative filtering online recommendation system

1. Online Movie
Recommendation System
(OMRES)
Sohel Kazi
03496202719
Hridik Gulati
04596202719

2. Content
• Introduction
• Problem Statement
• Background
• Proposed Solution
• Conclusion

3. Introduction
• People seek information via words,
recommendation letters, news reports etc.
• Recommendation systems imitate this social
process to enable quick filtering of the
information on the web
• Lots of companies try to offer services that
involve recommendations to address the
right user groups.

4. Problem Statement
● Providing related content out of relevant and
irrelevant collection of items to users of
online service providers.
● OMRES (Online Movie Recommendation
System) aims to recommend movies to users
based on movie overview, genres, cast and
crew.

5. Problem Statement
● Given a set of movies database giving similar
movies recommendations to the users based
on user movie input having maximum
similarities based on genre, cast and crew.
● Ex:- “If the user movie input is Avatar than
the recommendations would be similar to it
based on given criteria which could be i. e.
Aliens.”

6. Background (Dataset)
● Tmdb 5000 Movie dataset that provides
5000 movies details like budget, genres,
homepage, ID, keywords, popularity and
many more details.
● Tmdb 5000 credits dataset that provides
5000 movies details like title, cast and crew.
● Each movie is represented by a unique ID.

7. Problem Background
• Earlier systems implemented in 1990s.
o GroupLens (Usenet articles) [Resnick, 1997]
o Siteseer (Cross linking technical papers)[Resnick,
1997]
o Tapestry (email filtering) [Goldberg, 1992]
• Earlier solutions provided for users to rate the item.

8. Problem Background
• Item-to-item collaborative filtering (people who buy x also
buy y), an algorithm popularized by Amazon.com's
recommender system
• Last.fm recommends music based on a comparison of the
listening habits of similar users
• Facebook, MySpace, LinkedIn, and other social networks
use collaborative filtering to recommend new friends

9. Background For the Solution
• Content Based Filtering
o Selects items based on the correlation between
the content of the items and the user’s
preferences
o Compare user profile to content of each item
• Collaborative Filtering
o chooses items based on the correlation between
people with similar preferences.
o Rate items based on ratings of the users that
rated the same items

10. Background For the Solution
• There are 3 types of CF
• Memory-based
• Model-based
• Hybrid
o Memory based and Model-based
o Content Based and Collaborative Filtering

11. Proposed Solution

12. Proposed Solution

13. Proposed Solution

14. Proposed Solution

15. Proposed Solution

16. Conclusion
• Recommendation systems provide content
for us by taking user input movie and
finding the most similar movies based on
the content.
• Content-based Filtering is a widely used
solution for this problem which we make use
of in our project

17. References
● https://www.kaggle.com/datasets/tmdb/tmd
b-movie
metadata?select=tmdb_5000_movies.csv
● Nearest Neighbor Approach using Clustering
on the TMDB 5000 data.

18. Questions