As links among users are the channel for interchanging information inside social networks, trust is a widely accepted indicator or parameter to discover and analyze potential security threats. This work proposes a trust evaluation schema for social networks based on the reliability (degree of trust between users) and the strength (closeness of the ties between users) of both direct and indirect links.

The reliability degree is calculated by analyzing interactions (for example, comments and forwarding) for direct links and by processing topological information for indirect links. The strength degree is calculated from historical user interaction data (for direct and indirect links). Those two degrees are combined in a link trust matrix, which supports a principled approach to model strong and weak trust.

The paper is definitely well written and structured. Indeed, the problem is properly introduced and formulated with the support of clear contextual definitions. Despite its complexity, the schema is relatively easy to understand. The experiments on simulated datasets show interesting results that seem to outperform existing state-of-the-art methods.