A Trust Based Information Dissemination Model for Evaluating the Effect of Deceptive Data* Yi Hu, Zhichun Xiao, and Brajendra Panda Computer Science and Computer Engineering Department University of Arkansas, AR 72701 Email: {yihucd@gmail.com, xiaozc@gmail.com, bpanda@uark.edu} Abstract This research studies the problem of evaluating the effect of deceptive data based on the Information Flow Network and the Web of Trust. We present an Information Dissemination Model that illustrates the prerequisite for dissemination of information based on the subject and object trusts. To evaluate the effects of deceptive data accurately, we offer a quantitative model that is utilized to calculate to what extent the subjects in the information flow network are affected. The algorithm for evaluating the spread of deceptive data is provided and the time complexity of the algorithm is also analyzed. 1. Introduction Trust and shared interest are the building blocks for most relationships in human society. Deceptive actions and the associated risks affect not only individuals, but also group of people. For semantic web, e-commerce web site, and online virtual communities, trust is not only a fundamental building block but is also critical for their operational success. Although trust relationship in cyberspace can be built up more quickly and easily, it is more fragile. A rumor circulated in a grocery store of a small town may slowly affect local patrons, while the communication speed and the number of affected people for the same information in an online community are generally much higher. This research concentrates on modeling and analyzing the effect of deceptive action in an open rating system; especially studying the way deceptive data flow among subjects in the web of trust. The contribution of this research lies in that we offer a *This work was supported in part by US AFOSR under grant FA9550-04-1-0429 model on assessing the result of deceptive action and how the information flow network and web of trust together affect the spread of information. 2. Background and Motivation Existing researches on trust management that are related to this paper are trust aggregation and trust propagation. Trust aggregation concentrates on giving an estimated trust rating for an object based on trust ratings given by different subjects for this object. Currently, there are two different approaches. One approach [1, 2, 3] does not consider the preference of each individual in the web of trust. Another approach [4] tries to utilize the preference of each individual to calculate a trust rating that may be different for each person. With the help of trust propagation, it is possible for a user to have an estimated trust rating on another without prior interactions. For example, say Alice trusts Bob and Bob trusts Cathy, if Alice also trusts Cathy to some extent based on Bob and Cathy’s trust relationships, then we say there is trust propagation here. In real world, we often trust the opinions of our friends’ friends. R. Guha [5] has proposed the notion of Atomic Propagation of trust which includes Direct Propagation and Co-citation. R. Levien proposed a trust propagation model called Advogato [6, 7], that is based on the network flow theory. C. Ziegler and G. Lausen proposed a trust propagation model, AppleSeed [8], based on spreading activation models [9]. 3. The Model What is the prerequisite for deceptive information to flow from a malicious user to the final receiver(s)? Or putting more generally, what is the prerequisite for a piece of information to flow from the source to its destination? Although there are lots of researches on information flow network, most of these researches Proceedings of the 40th Hawaii International Conference on System Sciences - 2007 1 © 1530-1605/07 $20.00 2007 IEEE Proceedings of the 40th Annual Hawaii International Conference on System Sciences (HICSS'07) 0-7695-2755-8/07 $20.00 © 2007