1. A Structured Approach to Query Recommendation With Social Annotation Data 童薇 2010/12/3

2. Outline Motivation Challenges Approach Experimental Results Conclusions 2010/12/3

3. Outline Motivation Challenges Approach Experimental Results Conclusions 2010/12/3

4. Motivation Query Recommendation  Help users search  Improve the usability of search engines 2010/12/3

5. Recommend what? Existing Work  Search interests: stick to user’s search intent Anything Missing?  Exploratory Interests: some vague or delitescent interests  Unaware of until users are faced with one  May be provoked within a search session 2010/12/3 smartphones apple products nexus one mobilemeipod touch equivalent or highly related queries apple iphone

6. Is the existence of exploratory interest common and significant? Identified from search user behavior analysis  Make use of one-week log search data Verified byStatistical Tests(Log-likehood Ratio Test)  Analyzethe causality betweeninitial queries and consequ- ent queries Results  In 80.9% of cases: Clicks on search results indeed affect the formulation of the next queries  In 43.1% of cases: Users would issue different next queries if they clicked on different results 2010/12/3

7. Two different heading directions of Query Recommendation Emphasize search interests:  Help users easily refine their queries and find what they need more quickly  Enhance the “search-click-leave” behavior Focus on exploratory interests:  Attract more user clicks and make search and browse more closely integrated  Increase the staying time and advertisement revenue Recommendqueries to satisfy both search and exploratory interests of users simultaneously 2010/12/3 equivalent or highly related queries apple iphone mobilemeipod touch nexus one

8. Outline Motivation Challenges Our Approach Experimental Results Conclusions 2010/12/3

9. Challenges To leverage what kind of data resource? Search logs: Interactions between search users and search engines Social annotation data: Keywords according to the content of the pages “wisdom of crowds” 2010/12/3

10. Challenges To leverage what kind of data resource? How to present such recommendations to users? Refine queries Stimulate exploratory interests 2010/12/3 A Structured Approach to Query Recommendation With Social Annotation Data

11. Outline Motivation Challenges Approach Experimental Results Conclusions 2010/12/3

12. Approach Query Relation Graph  A one-mode graph withthe nodes representing all the uniquequeries and the edges capturing relationships between queries Structured Query Recommendation  Rankingusing Expected Hitting Time  Clustering with Modularity  Labeling each cluster with social tags 2010/12/3

13. Query RelationGraph Query Formulation Model 2010/12/3

14. Query RelationGraph Query Formulation Model 2010/12/3 3 5 4 2 3 1 2

15. Query RelationGraph Query Formulation Model Construction of Query RelationGraph 2010/12/3 3 2 1 3 1 2 1 2 3 5 43 1 2

16. Ranking with Hitting Time Apply a Markov randomwalk on the graph Employ hitting time as ameasure to rank queries  The expected number of steps beforenode j is visited starting from node i  The hiting time T is the first time that the random walk is at node j from the start node i:  The mean hitting time h(j|i) is the expectation of T under the condition 2010/12/3

17. Ranking with Hitting Time Apply a Markov randomwalk on the graph Employ hitting time as ameasure to rank queries  The expected number of steps beforenode j is visited starting from node i  Satisfies the following linear system 2010/12/3

18. Clustering withModularity Group the top k recommendations into clusters  It is natural to apply a graph clustering approach  Modularity function Note: In a network in which edges fall between vertices without regard for the communities they belong to,we would have 2010/12/3

19. Clustering withModularity Group the top k recommendations into clusters  It is natural to apply a graph clustering approach  Modularity function  Employ the fast unfolding algorithm to perform clustering 2010/12/3

20. Clustering withModularity Group the top k recommendations into clusters  It is natural to apply a graph clustering approach  Modularity function  Employ the fast unfolding algorithm to perform clustering Label each cluster explicitly with social tags The expected tag distribution given a query: The expected tag distribution under a cluster: 2010/12/3

21. Outline Motivation Challenges Approach Experimental Results Conclusions 2010/12/3

22. Experimental Results Data set  Query Logs: Spring 2006 Data Asset (Microsoft Research) 15 million records (from US users) sampled over one month in May, 2006 2.7 million unique queries and 4.2 million unique URLs  Social Annotation Data: Delicious data Over 167 million taggings sampled during October and November, 2008 0.83 million unique users, 57.8 unique URLs and 5.9 million unique tags  Query Relation Graph: 538, 547 query nodes Baseline Methods  BiHit: Hitting Time approach based on query logs (Mei et al. CIKM ’08)  TriList: list-based approach to query recommendation considering both search and exploratory interests  TriStrucutre: Our approach 2010/12/3

23. Examplesof RecommendationResults Query = espn BiHit espn magazine espn go espn news espn sports esonsports baseball news espn espn mlb sports news espn radio espn 103.3 espn cell phone espn baseball sports mobile espn espn hockey TriList espn radio espn news yahoo sports nba news cbs sportsline espn nba sports espn mlb espn sports sporting news scout sportsline sports illustrated bill simmons fox sports TriStructure [sports espn news] espn radio espn news espn nba espn mlb espn sports bill simmons [sports news scores] yahoo sports nba news cbs sportsline sports sporting news scout sportsline sports illustrated fox sports 2010/12/3

24. Examplesof RecommendationResults Query = 24 BiHit 24 season 5 24 series 24 on fox 24 fox fox 24 24 tv show tv show 24 24 hour fox television network fox broadcasting fox tv fox sports net fox sport ktvi 2 fox five news TriList fox 24 kiefer sutherland tv guide 24 tv show 24 fox jack bauer grey’s anatomy 24 on fox desperate housewives prison break 24 spoilers abc tv listings fox one tree hill TriStructure [tv 24 entertainment] fox 24 kiefer sutherland 24 tv show 24 fox jack bauer 24 on fox 24 spoilers [tv televisions entertainment] tv guide abc tv listings fox [tv television series] grey’s anatomy desperate housewives prison break one tree hill 2010/12/3

25. Manual Evaluation Comparison based on users’click behavior  A label tool to simulate the real search scenario  Label how likelihood the user would like to click (6-point scale)  Randomly sampled 300 queries, 9 human judges 2010/12/3

26. Distributions of Labeled Score over Recommendations Experimental Results (cont.) Overall Performance non-zero label score ➡ click Clicked Recommendation Number (CRN) Clicked Recommendation Score (CRS) Total Recommendation Score (TRS) Click Performance Comparison 2010/12/3

27. How Structure Helps  How the structured approach affects users’ click willingness  Click Entropy Experimental Results (cont.) The Average Click Entropy over Queries under the TriList and TriStructure Methods. 2010/12/3

28. How Structure Helps  How the structured approach affects users’ click patterns  Label Score Correlation Experimental Results (cont.) Correlation between the Average Label Scores on Same Recommendations for Queries. 2010/12/3

29. Outline Motivation Challenges Approach Experimental Results Conclusions 2010/12/3

30. Conclusions Recommend queries in a structured way for better satisfying both search and exploratory interests of users Introduce the social annotation data as an important resource for recommendation Better satisfy users interests and significantly enhance user’s click behavior on recommendations Future work  Trade-off between diversity and concentration  Tag propagation 2010/12/3

31. Thanks ！ 2010/12/3