چكيده به لاتين
Social networks are location for sharing and exchanging information. Also, information are generated and consumed by users. Unfortunately, social networks are open environment and no central entity controls them. So, anyone can produce and publish information. On the other hand, due to the large size and dynamism of these networks each user knows only som small parts of other users and the remaining parts are unknown for him. So, the major parts of the interactions take place between strangers. Thus, the reliability assessment prior to interact with them is a fundamental problem in these networks. How to calculate the trust score between users who previously had no interaction with each other is one of the challenging issues in trust field. The algorithm, which inferences trust score of one person to another person who directly or indirectly connected to him, is called trust inference algorithm. Recent researches inference trust score of an anonymous user from the perspective of user who has an intention to interact with him.
In this thesis, a trust inference model which is based on users ranking and trust propagation features is presented. In this model, the rank or importance and effect of each node from the perspective of trust relations is calculated using the network structural information. Then, the score of unknown trust can be predicted by trust propagation trough trust relations. To evaluate the proposed method for predicting trust, three subsets of Ciao data set are used. The results are then compared with the three considered basic methods. According to the results, the presented method is more accurate than the other three considered methods.
