On computational approaches to trust evaluation in large-scale social networks

In this paper, we consider the problem of evaluating trust between individuals in social networks. The previously introduced three-valued subjective logic (3VSL) provides a set of useful tools in measuring interpersonal trust in social human networks. However, the number of required operations for such measurements grows exponentially with the size of network. Moreover, the correlations between different paths connecting the two individuals caused by common edges are not considered. In this paper, we show that the operators in 3VSL can still be used to give a lower-bound on trust even if such correlations are taken into account. We introduce a low complexity scalable algorithm to obtain this lower-bound. The numerical experiment results are represented and compared with the 3VSL.