![• Need to know about item content
– requires manual or automatic indexing
– Item features do not capture everything
Limitations of Traditional
Recommendation system:
8/23/2015 14
• Lack of serendipity
[Wikipedia: “the effect by which one accidentally
discovers something fortunate, especially while
looking for something entirely unrelated” ]](https://image.slidesharecdn.com/b7ppt-150823181449-lva1-app6892/85/B7-ppt-14-320.jpg)
B7 ppt
- 1. Recommendation System for Books 8/23/2015 1 Under the Guidance of Prof. Praveen M D Name USN Abhishek M G 2BV12IS002 Rashmi N T 2BV12IS047 Sachin Patil 2BV12IS077 Rishabh Mehta 2BV12IS071
- 2. 8/23/2015 2
- 3. Function Means Tree 8/23/2015 3 Decision tree matrix Euclidean distance
- 4. Dataset • 278,858 users • 1,149,780 ratings • 271,379 books 8/23/2015 4
- 5. Why KNN? • Non parametric • Decision based on complete data kNN Versus SVM in the Collaborative Filtering Framework Miha Grcar Jozef Stefan Institute Jamova 391000 Ljubljana, Slovenia miha.grcar@ijs.si 8/23/2015 5
- 6. Contents • Problem Statement • Introduction • Existing System • System Design • OFCM • System Evaluation • Course relevance • References 8/23/2015 6
- 7. To recommend top-N most relevant books for a user, using item based collaborative filtering & user based collaborative filtering techniques and evaluating the performance of these two techniques. 8/23/2015 7 Problem Statement
- 9. Are they effective?? (Celma & Lamere, ISMIR 2007) Netflix 2/3 rated movies are from recommendation Google News 38% more click-through are due tommendation Amazon 35% sales are from recommendation 8/23/2015 9 Introduction (Contd..)
- 10. 8/23/2015 A recommendation system... how its work? 10 Introduction (Contd..)
- 11. 8/23/2015 Recommender system (RS) help users find items (e.g., news items, movies,Books) that meet their specific needs. 11 Introduction (Contd..)
- 12. 8/23/2015 2 Common Approaches : 1.Collaborative filtering 2.Content-based filtering 12 Recommendation System
- 13. 8/23/2015 Based on a description of the item and a profile of the user’s preference (Brusilovsky Peter , 2007) Content Based Filtering 13
- 14. • Need to know about item content – requires manual or automatic indexing – Item features do not capture everything Limitations of Traditional Recommendation system: 8/23/2015 14 • Lack of serendipity [Wikipedia: “the effect by which one accidentally discovers something fortunate, especially while looking for something entirely unrelated” ]
- 15. 8/23/2015 A method of making automatic predictions (filtering) about the interests of a user by collecting preferences or taste information from many users (collaborating) Collaborative Filtering 15
- 16. 8/23/2015 16 System Design Recommendation System for Books
- 17. User-based collaborative filtering Item-based collaborative filtering Types Of Collaborative Filtering 8/23/2015 17
- 30. 30 Item Similarity Calculation Adjusted Cosine Similarity 8/23/2015
- 31. 31 Item Similarity Calculation Example 1 2 3 4 5 6 1 2 3 4 5 6 8/23/2015
- 32. 32 Item Similarity Calculation 1 1 2 3 4 5 6 8/23/2015
- 33. 33 Similar Item 1 4 3 6 5 2 8/23/2015
- 35. • MAE—Mean Absolute Error • RMSE--Root mean squared error System Evaluation 8/23/2015 35
- 36. 8/23/2015 36 Course Relevance Sl No : Concept Used Subject 1 Data preprocessing, classification, Prediction. Data Mining, Machine Learning 2 PHP,HTML5 Web Technology 3 Similarity Computation. Linear Algebra 4 Collaborative filtering algorithms D.S & Algorithms
- 37. References FRANCESCO RICCI .(2010). Recommender Systems Handbook. LONDON:SPRINGER. ALA ALLUHAIDAN. Recommender System Using Collaborative Filtering Algorithm. JOONSEOK LEE.(2012).A Comparative Study of Collaborative Filtering Algorithms. B.M. Sarwar et al., “Item-Based Collaborative Filtering RecommendationAlgorithms,” 10th Int’l World Wide Web Conference, ACM Press, 2001, pp. 285-295. 8/23/2015 37
- 38. Thank You ! 8/23/2015 38
- 39. 8/23/2015 39