Classifying trust/distrust relationships in online social networks

Giacomo Bachi, Michele Coscia, Anna Monreale, Fosca Giannotti

Research output: Contribution in Book/Catalog/Report/Conference proceedingConference contribution

Abstract

Online social networks are increasingly being used as places where communities gather to exchange information, form opinions, collaborate in response to events. An aspect of this information exchange is how to determine if a source of social information can be trusted or not. Data mining literature addresses this problem. However, if usually employs social balance theories, by looking at small structures in complex networks known as triangles. This has proven effective in some cases, but it under performs in the lack of context information about the relation and in more complex interactive structures. In this paper we address the problem of creating a framework for the trust inference, able to infer the trust/distrust relationships in those relational environments that cannot be described by using the classical social balance theory. We do so by decomposing a trust network in its ego network components and mining on this ego network set the trust relationships, extending a well known graph mining algorithm. We test our framework on three public datasets describing trust relationships in the real world (from the social media Epinions, Slash dot and Wikipedia) and confronting our results with the trust inference state of the art, showing better performances where the social balance theory fails.

Original languageEnglish
Title of host publicationProceedings - 2012 ASE/IEEE International Conference on Privacy, Security, Risk and Trust and 2012 ASE/IEEE International Conference on Social Computing, SocialCom/PASSAT 2012
Pages552-557
Number of pages6
DOIs
Publication statusPublished - 2012
Externally publishedYes
Event2012 ASE/IEEE International Conference on Social Computing, SocialCom 2012 and the 2012 ASE/IEEE International Conference on Privacy, Security, Risk and Trust, PASSAT 2012 - Amsterdam, Netherlands
Duration: 3 Sept 20125 Sept 2012

Conference

Conference2012 ASE/IEEE International Conference on Social Computing, SocialCom 2012 and the 2012 ASE/IEEE International Conference on Privacy, Security, Risk and Trust, PASSAT 2012
Country/TerritoryNetherlands
CityAmsterdam
Period3/09/125/09/12

Keywords

  • graph mining
  • mining
  • social networks
  • trust

Fingerprint

Dive into the research topics of 'Classifying trust/distrust relationships in online social networks'. Together they form a unique fingerprint.

Cite this