A Computational Trust Model for E-Commerce Systems: Concepts, Definitions and Evaluation Method

Trust is a complex and multidimensional concept, which plays a key role in the success of electronic commerce. Assessing trust, specifically in the beginning of a commercial relation and the formulation of trust in general is a complex and difficult task. The researchers are often focused on a specific context for trust formulation and the relevant literature does not clearly distinguish between the factors involving in trust decision making process. With the aim of providing a basis for computational trust models and by consolidating a large body of studied contexts in the trust literature, this paper first tries to present a conceptual trust model for electronic commerce. Four types of trust that are used in the conceptual trust model are as follows: (۱) institutional trust, (۲) technological trust, (۳) trading party trust, and (۴) propensity trust. Then, a computational trust model is proposed in which the agents involved in a commercial transaction can consult with a trust manager agent (TMA), which is considered in a distributed fashion in the network. The proposed model is capable of evaluating a broad range of trust contexts and has two main features: (۱) trust is evaluated dynamically (i.e., a change in any of the trust’s parameters will result in the re-calculation of trust values) and (۲) the proposed model is capable of making partial studies for the trust contexts presented in the conceptual model of trust. Finally, the proposed model is evaluated and the results are presented in this paper.

Keywords:

Trust model , trust evaluation , computational trust model , electronic commerce , trust manager agent (TMA)