Trust reality-mining: evidencing the role of friendship for trust diffusion

• Published in: Human-centric computing and information sciences 2017-02, Vol.7 (1), p.1, Article 4
• Main Authors: Farrahi, Katayoun, Zia, Kashif
• Format: Article
• Language: English
• Subjects: Artificial Intelligence; Communications Engineering; Computation; Computer Science; Computer Systems Organization and Communication Networks; Data analysis; Identification methods; Information Systems and Communication Service; Information Systems Applications (incl.Internet); Information technology; Networks; User Interfaces and Human Computer Interaction
• ISSN: 2192-1962; 2192-1962
• DOI: 10.1186/s13673-016-0085-y

Value sensitive design is driven by the motivation of making social and moral values central to the development of ICT systems. Among the most challenging concerns when imparting shared values like accountability, transparency, liberty, fairness and trust into information technology are reliable and comprehensive formal and computational models of those values. This paper, educated by trust theories and models from cognitive science, social sciences and artificial intelligence, proposes a novel stochastic computational model of trust, encapsulating abstractions of human cognitive capabilities and empirically evidenced social interaction patterns. Qualitative and quantitative features of trust are identified, upon which our formal model is phrased. Reality mining methods are used to validate the model based on a real life community dataset. We analyze the time-varying dynamics of the interaction and communication patterns of the community, consider varying types of relationships as well as their symmetry. Social network data analysis shows that our model better fits the evolved friendships compared to a well designed synthetic trust model, which is used as the baseline.