recommender systems
TRANSCRIPT
![Page 1: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/1.jpg)
Recommender Systems
Anastasiia Kornilova
![Page 2: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/2.jpg)
Agenda
Problems
Evaluation
Algorithms
General overview
![Page 3: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/3.jpg)
We are overloaded of information:
• Books • Movies • News • Blogs • TV-channels • Music • …
![Page 4: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/4.jpg)
As user:
• Do we need all of this things?
• No
• Can we choose most appropriate of them?
• Yes
• How?
• Recommender systems!
![Page 5: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/5.jpg)
As Business Owner:
Do you need Recommender System?
• Netflix:
• 2/3 rented movies are from recommendation
• Google News
• 38 % more click-through are due to recommendation
• Amazon
• 35% sales are from recommendation
•Celma & Lamere, ISMIR 2007
![Page 6: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/6.jpg)
Domains of recommendations
Content to Commerce
• Information
• News
• Restaurants
• Vendors
• TV-programs
• Courses in e-learning
• People
• Music playlists
One particularly interesting
property
• New items (movies, books, news, ..)
• Re-recommend old ones (groceries, music,…)
![Page 7: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/7.jpg)
Examples: Retail
![Page 8: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/8.jpg)
Examples: Banking
Deposits:
Deposit 1
Deposit 2
Credit products:
Credit card 1
Credit card 2
Personal loan 1
Personal loan 2
Insurance:
Endowment
Travel insurance
Service packages:
Premium package
Customer 2
Premium package
Travel insurance
Credit card 1
Customer 1
Travel insurance
Personal load 2
Deposit 1
Consultant can be replaced by Recommender System
![Page 9: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/9.jpg)
Examples: Hotels
![Page 10: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/10.jpg)
Examples: Advertisement
Shopping (browsing) history RSE
![Page 11: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/11.jpg)
Purposes of Recommendation
Recommendations themselves (Sales, information)
Education of user/customer
Build a community of users/customers around products or content
![Page 12: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/12.jpg)
Whose Opinion?
“Experts”
Ordinary “phoaks”
People like you
![Page 13: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/13.jpg)
Personalization Level
Generic/Non-Personalized: everyone receives same recommendations
Demographic: matches a target group
Ephemeral: matches current activity
Persistent: matches long-term interests
![Page 14: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/14.jpg)
Explicit input based RSE
Rating
Review Vote
Like
![Page 15: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/15.jpg)
Implicit input
based RSE
Click
Purchase
Follow
![Page 16: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/16.jpg)
Recommendation Algorithms
1. Non-Personalized Summary Statistics
2. Content-Based Filtering
Information Filtering
Knowledge-Based
3. Collaborative Filtering
User-user
Item-item
Dimensionality Reduction
4. Others
Critique / Interview Based Recommendations
Hybrid Techniques
![Page 17: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/17.jpg)
Non-Personalized Recommender
Best-seller
Most popular
Trending Hot
Best-liked
People who X also Y
![Page 18: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/18.jpg)
Personalized Recommender:
Collaborative Filtering
Use opinions of others to predict/recommend
User model – set of ratings
Item model – set of ratings
Common core: sparse matrix of ratings
![Page 19: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/19.jpg)
Evaluation
Lift Cross-sales Up-sales
Conversions Accuracy Serendipity
![Page 20: Recommender systems](https://reader034.vdocuments.site/reader034/viewer/2022042816/5587b3a9d8b42a61408b4752/html5/thumbnails/20.jpg)
Problems
“Cold start”
New user
New item
New system