Sas web 2010 lora-aroyo
0 likesâą362 views
This document proposes a model for representing trust and reputation in social internetworking systems. It discusses representing users, resources, and their interactions as a heterogeneous hypergraph rather than a traditional social network graph. It also presents algorithms for computing trust and reputation through a mutual reinforcement principle between user reputation and resource quality ratings. Future work is outlined to test these approaches on real social networking data and domains.
Sas web 2010 lora-aroyo
- 1. Trust and Reputation in Social Internetworking Systems Lora Aroyo1 Pasquale De Meo1 Domenico Ursino2 1VU University Amsterdam, the Netherlands 2DIMET â University of Reggio Calabria, Italy
- 2. Social Networks Added Value ! â advertise products and disseminate innovations & knowledge ! â find information relevant to users ! â find relevant users, e.g. LinkedIn ! â spread opinions, e.g., personal, social or political ! â interesting for: ! â museums, broadcaster, government institutions
- 3. Online Identities ! â Increasing number of identities ! â different information sharing tasks ! â connect with different communities ! â UK adults have ~1.6 online profiles ! â 39% of those with one profile have at least two other profiles ! â Companies exploring the potential of social internetworking ! â Platform(s) for data portability among social networks
- 4. Social Internetworking System © danbri
- 5. Whatâs Needed? ! â mechanisms to: ! â help users find reliable users ! â disclose malicious users or spammers ! â stimulate the level of user participation ! â deal with trust in linked data ! â deal with different contexts and policies for accessing, publishing and re-distributing data
- 6. Whatâs the Goal? ! â model to represent Social Internetworking components & their relationships ! â understand Social Internetworking structural properties and see how it differs from traditional social networks ! â model to compute trust & reputation based on linked data
- 7. Requirements ! â trust should be tied to userâs performance, i.e., providing beneficial contributions to other users ! â consider that users are involved in a range of activities, e.g., tagging, posting comments, rating ! â represent a wide range of heterogeneous entities, e.g. users, resources, posts, comments, ratings and their interactions (vs. single role nodes in graphs) ! â edges need to support n-ary relationships vs. binary in graphs ! â multi-dimensional network vs. one-dimensional in graphs ! â easy to manipulate and intuitive model
- 8. Graph-Based Approaches ! â Model user community as graph G ! â edges reflect explicit trust relationship between users ! â G is sparse, thus often need for inferring trust values ! â model trust & reputation in force-mass- acceleration style ï capture all factors and combine them in a set of equations ! â resulting model is too complicated to be handled
- 9. Link-Based Approaches ! â link analysis algorithms, e.g. PageRank or HITS, model trust as a measure of system performance, e.g., number of corrupted files in a peer of a P2P network ! â attack-resistant to manipulate reputation score ! â model trust & reputation in force-mass- acceleration style ï capture all factors and combine them in a set of equations ! â resulting model is too complicated to be handled
- 10. SIS Approach ! â Social Graph API (list of public URLs and connections for person p (e.g., Twitter page of p and contacts of p) ! â Hypergraph ! â nodes labels with object role ! â multiple hyperedges between two nodes ! â hyperedges â link two or more entities
- 11. SIS Pilot: Analysis ! â We gathered from multiple social networks, e.g., LiveJournal, Twitter, Flickr: ! â 1, 252, 908 user accounts ! â 30, 837, 012 connections between users ! â The probability P(k) that a user has created an account in k networks is distributed as: P(k) ~ k-4.003 ! â Few users are affiliated to multiple networks ! â More than 90% of users are affiliated to less than 3 networks
- 12. Canonization Procedures ! â Map gathered data to graph with following properties: ! â High network modularity, i.e., nodes tend to form dense clusters with few inter-cluster edges ! â Small world phenomenon, i.e., paths between arbitrary pairs of nodes are usually short
- 13. Reputation in SIS ! â Setting: ! â users post resources &rate resources posted by others ! â To compute reputation we assume that: ! â User-high-reputation if he authors high quality resources ! â Resource-high-quality if it gets a high average rating & posted by users with high reputation ! â mutual reinforcement principle
- 14. Trust in SIS ! â n = # of users in SIS m = # of resources they authored ! â r(i) = reputation of useri q(j) = quality of resourcej ! â e(j) = average rating of resourcej ! â Aij = 1 if useri posted a resourcej and Aij = 0 otherwise ! â r = Aq and q = AT r + e ïš r = (I â AAT)-1Ae . ! â compute dominant eigenvector of a symmetric matrix ! â easy to compute even if A gets large (AT = transpose of A and I = nxn identity matrix)
- 15. Future Work ! â Gather a larger amount of data to analyze further the structural properties of SIS ! â Test the effectiveness of the approach for reputation computing ! â Test with real users in the social space of Agora (Social Event-based History browsing) and in PrestoPrime (Social Semantic Taging) . ! â Ontology-based model of trust and reputation in different domains (with LOD)
- 16. Acknowledgements ! â This research is funded by EU Marie Curie Fellowship Grant