Recommendation system | Group Presentation

"This presentation explores the fundamentals of recommendation systems, including collaborative filtering, content-based filtering, and hybrid approaches. Learn how AI and machine learning power personalized recommendations in e-commerce, streaming platforms, and other industries."

1. Recommendation system 1

2. What 2

3. Use-Cases Of Recommendation System Personalized Content Improve the on-site experience Better Product search experience categorization of products 3

4. Good Recommendation Recommends available products Relevant to that user Personalized doesn’t recommend the same items 4

5. TYPES OF RECOMMENDATION SYSTEM 5

6. PERSONALIZED 6

7. Content-Based Filtering 7

8. Read by User Similar books Recommended to user 8

9. Advantages since recommendations are specific to a single user doesn’t need data of other users It makes it easier to scale to a large number of users easier to scale The model can Capture the specific Interests of the user and can recommend items that very few other users are interested in. Capture the specific Interests 9

10. Disadvantages Feature representation of items is hand-engineered to some extent, this tech requires a lot of domain knowledge hand-engineered The model can only make recommendations based on the existing interest of a user. In other words, the model has limited ability to expand on the user’s existing interests limited ability 10

11. Collaborative Filtering 11

12. Types of collaborative filtering Memory based applied to raw data without any prior preparation Model based developed using machine learning algorithms 12

13. Memory based 13

14. Memory based User based 14

15. Memory based Item based 15

16. Advantages It works well even if the data is small Data small This model helps the users to discover a new interest in a given item but the model might still recommend it because similar users are interested in that item. discover a new interest No need for Domain Knowledge Domain Knowledge 16

17. Disadvantages Side Feature Doesn’t have much importance. Here Side features can be actor name or releasing year in the context of movie recommendation. Side Feature It cannot handle new items because the model doesn’t get trained on the newly added items in the database. This problem is known as Cold Start Problem Cold Start Problem 17

18. Model based 18

19. Matrix Factorization 19

20. 20

21. Hybrid 21

22. They represent a combination of different recommenders. The idea is that combining several different recommenders will produce better results than using a single algorithm. 22

23. Team Amit Golder 201-15-3058 Marcel David Baroi 201-15-3421 Shatabdi Shil Piu 201-15-3448 Tanvir Niaz Khan 201-15-3593 23

24. THANK YOU 24