A Trust Evaluation Model for Future Peer-to-peer E-Commerce Environments

Abstract

Due to the flexibility and freedom of the P2P e- commerce environment, peers often make transactions with other unknown peers, the trust evaluation becomes critical. Traditionally the trust evaluation model is based on feedback mechanism, which only depends on transactions number neglecting transaction amount, etc. This may lead to the transaction risk in the new transaction. In this paper, a novel trust evaluation model is presented, which reflects recommender's trust value and transaction amounts when computing the trust value of peer. In addition, the model introduces the factor of transaction period while updating peer's trust value, which embodies that the feedback occurred in different period have different influence to peer's trust value. Therefore results of trust evaluation become more accurate, the novel trust evaluation model is effective against malicious behavior. In P2P e-commerce environment, it lacks the central management. Peers are unfamiliar each other. It is rational to doubt the other peer's trustworthiness. The environment must have an effective trust system to help users locate trustworthy partners and exchange commodities securely with confidence. The trust evaluation model based on feedbacks give a solution to the above-mentioned problem. When a peer wants to make transaction with the unknown destination peer, the peer can enquire other peers about the destination peer's transaction trust feedbacks. After having collected the feedbacks from a set of responding peers, the requesting peer can analyze the data and evaluate the trust status of the destination peer. In a certain extent, the trust evaluation model is helpful to evaluate the peer's trust status. When the trust status of the destination peer is evaluated, other peers' feedbacks outline the transaction history of the destination peer providing an indication of the transaction trust status of the destination peer, which in some sense indicates the risk level of the new transaction. However the possible new transaction may be different from some previous transactions, for example transaction amount, which implies these old transactions should not be taken as references equally when analyzing the transaction trust of the destination peer. Otherwise, it may lead to a result with bias. This is especially risky for a -peer, if it has no transaction previously with the destination peer. However, most existing trust evaluation models for P2P e- commerce obtain the peer's trust status only considered by transaction number. But other factors are not taken into account, such as transaction amount, the trust value of peer, etc. This paper presents a novel trust evaluation model, which is based on the peer's transaction experience �» but distinguishes transaction amounts, and introduces impact factors of transaction period and the recommender's trust value. Therefore the result of trust evaluation becomes more accurate, the obtained trust status of peer can be taken as the risk indication of the forthcoming transaction.