The Ripple Effect: Change in Trust and Its Impact Over a Social Network

Computing trust between individuals in social networks is important for many intelligent systems that take advantage of reasoning in social situations. There have been many algorithms developed for inferring trust relationships in a variety of ways. These algorithms all work on a snapshot of the network; that is, they do not take into account changes in trust values over time. However, trust between people is always changing in realistic social networks and when changes happen, inferred trust values in the network will also change. Under these circumstances, the behavior the existing trust-inference algorithms is not yet very well understood. In this paper, we present an experimental study of several types of trust inference algorithms to answer the following questions on trust and change: How far does a single change propagate through the network? How large is the impact of that change? How does this relate to the type of inference algorithm? Our experimental results provide insights into which algorithms are most suitable for certain applications.