RecommendationSystems-2

RecommendationSystems-2 - CS345 Data Mining Recommendation...

Info iconThis preview shows pages 1–13. Sign up to view the full content.

View Full Document Right Arrow Icon
    CS345 Data Mining Recommendation Systems Anand Rajaraman, Jeffrey D. Ullman
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Recommendations  Items Search Recommendations Products, web sites, blogs, news items, …
Background image of page 2
The Long Tail Source: Chris Anderson (2004)
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
From scarcity to abundance Shelf space is a scarce commodity for  traditional retailers  Also: TV networks, movie theaters,… The web enables near-zero-cost  dissemination of information about products From scarcity to abundance More choice necessitates better filters Recommendation engines How  Into Thin Air  made  Touching the Void  a  bestseller
Background image of page 4
Recommendation Types Editorial Simple aggregates Top 10, Most Popular, Recent Uploads Tailored to individual users Amazon, Netflix, …
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Formal Model C  = set of Customers S  = set of Items Utility function  u C   £ S   !   R = set of ratings R  is a totally ordered set e.g., 0-5 stars, real number in [0,1]
Background image of page 6
Utility Matrix 0.4 1 0.2 0.3 0.5 0.2 1 King Kong King Kong LOTR LOTR Matrix Matrix National Treasure National Treasure Alice Alice Bob Bob Carol Carol David David
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Key Problems Gathering “known” ratings for matrix Extrapolate unknown ratings from known  ratings Mainly interested in high unknown ratings Evaluating extrapolation methods
Background image of page 8
Gathering Ratings Explicit Ask people to rate items Doesn’t work well in practice – people can’t be  bothered Implicit Learn ratings from user actions e.g., purchase implies high rating What about low ratings?
Background image of page 9

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Extrapolating Utilities Key problem: matrix U is sparse most people have not rated most items Three approaches Content-based Collaborative Hybrid
Background image of page 10
Content-based recommendations Main idea: recommend items to customer C  similar to previous items rated highly by C Movie recommendations recommend movies with same actor(s), director,  genre, … Websites, blogs, news recommend other sites with “similar” content
Background image of page 11

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
likes likes Item profiles Item profiles Red
Background image of page 12
Image of page 13
This is the end of the preview. Sign up to access the rest of the document.

This document was uploaded on 03/04/2012.

Page1 / 35

RecommendationSystems-2 - CS345 Data Mining Recommendation...

This preview shows document pages 1 - 13. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online