It is well known that the fundamental intellectual problems of information access are the production and consumption of information. In this paper, we investigate the use of social network of information producers (authors) within relations in data (co-authorship and citation) in order to improve the relevance of information access. Relevance is derived from the network by levraging the usual topical similarity between the query and the document with the target author’s authority. We explore various social network based measures for computing social information importance and show how this kind of contextual information can be incorporated within an information access model. We experiment with a collection issued from SIGIR proceedings and show that combining topical, author and citation based evidences can significantly improve retrieval access precision, measured in terms of mean reciprocal rank.
Driving Behavioral Change for Information Management through Data-Driven Gree...
An Exploratory Study on Using Social Information Networks for Flexible Literature Access
1. An exploratory study on using social informationnetworks for flexible literature access Lynda Tamine, Amjed Ben Jabeur and WahibaBahsoun University Paul Sabatier Toulouse III, France IRIT SIG-RI {lechani, jabeur,wbahsoun}@irit.fr FQAS 2009
2. An exploratory study on using social information networks for flexible literature access Outline: Social Information Retrieval : Background and motivation A social based model for literature access Experimental evaluation Conclusion and outlook 2
3. Towards Social Information Retrieval IRS Query User Profile Tag Comment 3 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
4. Towards Social Information Retrieval Information Producer Documents Social Information Retrieval Queries Information consumer Tags, comments ..etc 4 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
5. About Social Information Retrieval Incorporating information from social network in the information retrieval process Access to relevant information in the social neighborhood Spread information through the social network Take account of the social activity Crossing two domains [KIRSCHet al , 2003] Information Retrieval Represent and compare document /query Social Network Analysis [WESSERMANN & FAUST] Represent social entities and relationships Estimate individual’s centrality 5 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
6. About Social Information Retrieval Social relevance features [AMER & al, 07] Topical relevance Social distance Incoming links and bookmarks Timeliness and freshness Social importance of individuals Relevant documents are published by important people 6 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
7. About Social Information Retrieval Domain application of Social Information Retrieval Knowledge and experience sharing , Collaborative production [KORFAITIS & al, 2006] Wiki, SourgeForge Opinion retrieval [ZANG & YE, 2008] Blogs Social media and bookmarking [HEYMANN & al, 2008] [BUDURA & al, 2008] Facebook, YouTube, Del.ici.us Information finding and social network exploring [ZANG & al , 2008] Social ranking of document, Expert searching Literature access [KIRSCH & al , 2006] CiteULike.org 7 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
8.
9. Related works Extract social network form bibliographic resources [KIRSCH, 2003] [MUTSCHKE, 2001] [KIRCHNOFF & al, 2008] Actors : documents and authors Edges : co-author relationships Multiplicative relevance score [KIRSCH, 2003] Query document similarity Author’s authority 9 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
10. Contribution Relationships extracted from bibliographic resources : Co-authorship Citation link Linear combination Topical relevance Social importance of documents Study the impact of centrality measures 10 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
11. Social Content Graph a1 a3 a2 a4 a6 a1 a3 a4 a2 a5 Co-author d1 d2 d3 d4 Citation Social network graph V: Authors E: Relationships between authors An edge express: Co-auteur relationship (implicit direct) Citation relationship (implicit indirect) 11 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
12. Estimate document relevance Step 1. Social Importance of authors a1 a3 a4 a2 Ranked List Social Importance Algorithms Of authors Degree, Closeness, Betweeness PageRank and Hits {author, score} Social Network Social Network Analysis Social Scores 12 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
20. Weighted Sum, Min, Max, AVG …etcImp(d) C(v4) C(v3) 14 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
21. Estimate document relevance Step 3. Retrieving documents Step 4. Combining topical relevance and social importance Result set IRS Query {d, RSV (q,d)} Topical relevance Social importance 15 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
22. ACM SIGIR 1978-2008 Metadata gathered from ACM Portal Author Citations links Downloads ( Mars 2008 – Mars 2009) Experimentation Degree distribution Verticeswithdegreeδ(ν) Vertex degreeδ(ν) 16 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
32. -------------1 {titleterms} Most downloaded assumption 2 Known item retrieval { d,q } 50 x{ d, q, rank } 50 x { document, query} 17 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
33. Comparing importance measures Most cited assumption Most downloaded assumption Co-author Citation Co-author & Citation 18 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
34. Parameter Tuning Most cited assumption PageRank PageRank HITS Closeness Most downloaded assumption Co-author & Citation Co-author Co-author Citation Co-author & Citation Citation 19 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
35. Comparative evaluation Co-author 27% 0.212 0.270 26% 59% Most cited 27% 61% Most cited Citation 59% 59% Co-author & Citation 61% 61% 20 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
36. Conclusion A social network based information access model Co-author and citation relationships Linear combination of scores Study effectiveness of the model using several social importance measure Ensure the soundness of our results using two social relevance assumptions 21 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
37. Outlook Extend the social content graph Additional relationships Weighting social relationships Include tag entity Test our retrieval model on a large web collection Study recall and precision of the proposed approach 22 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
39. References [KIRSH & al, 06] Sebastian Kirsch, Melanie Gnasa, Armin Cremers, “Beyond the Web : Retrieval in Social Information Spaces” , Proceedings of the 28th European Conference on Information Retrieval, ECIR 2006, Imperial College, London, 2006 [AMER & al, 07] SihemAmer-Yahia, Michael Benedikt,Philip Bohannon, “Challenges in Searching Online Communities” , IEEE Data Eng. Bull., 2007 [KIRCHNOFF & al, 08] Lars Kirchhoff, Katarina Stanoevska-Slabeva, Thomas Nicolai and Matthes Fleck , “Using social network analysis to enhance information retrieval systems , Applications of Social Network Analysis”, ASNA, Zurich, 2008 [LANGVILLE & MEYR, 08] Lars Kirchhoff, Katarina Stanoevska-Slabeva, “Using social network analysis to enhance information retrieval systems , Applications of Social Network Analysis”, ASNA, Zurich, 2008 24 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
40. References [MUTSCHKE, 01] Peter Mutschke, “Enhancing Information Retrieval in Federated Bibliographic Data Sources Using Author Network Based Stra-tagems”, Reserach and Advanced Technology for Digital Libraries : 5th European Conference, ECDL 2001, Darmstadt, Germany, September 4-9, 2001 [EDGAR & RIJIKE, 07] Edgar Meij and Maarten de Rijke, “Using Prior In-formation Derived from Citations in Literature Search” Proceedings of RIAO 2007 : Recherched’InformationAssistée par Ordinateur 2007, 2007 [KORFIATIS & al., 2006] Korfiatis, N.; Poulos, M. & Bokos, G. “EvaluatingAuthoritative Sources using Social Networks: An Insight fromWikipedia”, Online Information Review, 2006, 30, 252-262 [HEYMANN & al, 2008] Paul Heymann, Daniel Ramage, Hector G. Molina "Social tag prediction Export" In SIGIR '08: Proceedings of the 31st annual international ACM SIGIR conference on Research and development in information retrieval (2008), pp. 531-538. 25 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
41. References [BUDURA et al.] Adriana Budura, Sebastian Michel, Philippe C. Mauroux, Karl Aberer “ To tag or not to tag -: harvesting adjacent metadata in large-scaletaggingsystemsExpor ”, In SIGIR '08: Proceedings of the 31st annual international ACM SIGIR conference on Research and development in information retrieval (2008), pp. 733-734. [ZHANG et al.] Jing Zhang, Jie Tang and Juanzi Li “ Expert Finding in a Social Network ”, Advances in Databases: Concepts, Systems and Applications, Volume 4443/2008, 1066-1069, Springer Berlin / Heidelberg, 2008 [ZANG & YE, 2008 ] Min Zhang and XingyoaYe “ A generation model to unifytopic relevance and lexicon-based sentiment for opinion retrieval ”, In Proceedings of the 31st Annual international ACM SIGIR Conference on Research and Development in information Retrieval (Singapore, Singapore, July 20 - 24, 2008). SIGIR '08. ACM, New York, NY, 411-418 26 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook
42. References [WASSERMAN & KHATERINE] Stanley Wasserman and Katherine Faust “ Social network analysis: methods and applications”, Cambridge Uni. Press [KIRSCH, 2003] Sebastian Marius Kirsch, “ Social information retrieval ”, PhDThesis in Computer Science, Computer science department III, Bonn, 14 March 2003 [KIRSCH & al. , 2006] Sebastian Kirsch, MelanieGnasa, and Armin Cremers “ Beyondthe web: Retrieval in social information spaces”, EuropeanConference on IR Research No28, London , ROYAUME-UNI (2006) 2006 [RITCHIE & TEUFEL, 2007] Anna Ritchie, Simone Teufel and Stephen Robertson, “ Usingtermsfrom citations for IR: some first results”, Proceedings of the EuropeanConference for Information Retrieval ECIR, pp 211-221, 2007 [SMALL, 1973] Henry Small “ Co-citation in the scientificliterature: A new measurement of the relationshipbetweentwodocuments”, Journal of the American Society of Information 27 Background and motivation | Social based model for literature access | Experimental evaluation | Conclusion and outlook