Abstract:

A peer-to-peer (P2p) trust framework is being developed that takes into account not only directly verifiable information, but also, as its new contribution, information that could be regarded as word-of-mouth advertising or hearsay. By supplying this sort of information, it is intended to both broaden the available knowledge of nodes at the fringes of the networks - i.e. not-highly connected entities, such as a casual online shopper - and explicitly expand existing trust models to further mimic real-life social behavior. Furthermore, dynamic effects of complex networks in the dissemination of reputation information for trust-decision making in social networks are leveraged. A formalization for the inclusion of so-called 'weak reputation' is presented, and prototypically implemented through the use of fuzzy logic mechanisms. Preliminary simulation results showing the general practicability of the concept are presented.