A strategy for trust propagation along the more trusted paths

The main goal of social networks are sharing and exchanging information among users. With the rapid growth of social networks on the Web, the most of interactions are conducted among unknown individuals. On the other hand, with increasing the biased behaviors in online communities, ability to assess the level of trustworthiness of a person before interacting with him has an important influence on users' decisions. Trust inference is a method used for this purpose. This paper studies propagating trust values along trust relationships in order to estimate the reliability of an anonymous person from the point of view of the user who intends to trust him/her. It describes a new approach for predicting trust values in social networks. The proposed method selects the most reliable trust paths from a source node to a destination node. In order to select the optimal paths, a new relation for calculating trustable coefficient based on previous performance of users in the social network is proposed. In ciao dataset there is a column called helpfulness. Helpfulness values represent previous performance of users in the social network. Advantages of this algorithm is its simplicity in trust calculation, using a new entity in dataset and its improvement in accuracy. The results of the experiments on Ciao dataset indicate that accuracy of the proposed method in evaluating trust values is higher than well-known methods in this area including TidalTrust, MoleTrust methods.