TP-TA: a comparative analytical framework for trust prediction models in online social networks based on trust aspects

Formation of online social network (OSN) strongly depends on the quality of relationships between its agents. Such relationships are affected by a host of factors; trust is one of them. To enhance the quality of relationships in such networks, it is important to find a mechanism to predict the degree of trust among participating agents since trust is the major driving force for initiating and developing social relationships. Although much effort has been made to develop quantitative techniques to obtain trust value, there is a lack of coherent classification of such techniques to achieve a macro vision of trust prediction models and identify their strengths and weaknesses. In this paper, we proposed TP-TA, an analytical framework which consists of three main components: First, classification of various existing trust prediction models in terms of trust aspects in the context of OSNs. Besides, main ideas, prospects, and challenges of each approach are highlighted for further research in this field. Second, defining general criteria to analyze our proposed classification. Finally, we illustrate a qualitative comparison between each approach which is a guide to understanding their superiority to one another. This framework could lead to an efficient selection of trust prediction techniques based on the nature of the target OSN and the intended trust type.