Study On Construction And Core Algorithms Of Reputation-Based Integrated Trust Model For Online Transaction

Security is a key issue in Online Transaction, traders'reputation evaluation and trust management are the foundation and difficulty to transaction security. The precondition of transaction is firstly to establish trust relationship through reputation evaluation to counterparts. Both sides mutual trust and degree of trust will determine whether transaction relationship occurrence, and trasaction investment level. Camparing to the directly face-to-face transaction, Reputation evaluation is very important to network transaction parties, because the compositon of network transaction parties is complex, and the fictitiousness and anonymous characteristics in network transactionm, which gets the Speculation Behavior often occurrence. So which must improve Online Transaction reliability, and reduce transaction risk? On this background,there appeares Reputation Evaluation and Trust Management System.The existence of malicious traders could damage the correctness and availability of the reputation systems. Reputation-Based trust mechanisms can recognize these malicious traders by computing the trustworthiness of the traders. The validity of a reputation-based trust model relies on some well-chosen trust factors, which directly influence the computation of trust value, the accuracy of the trust model, and the resistibility of the trust model to various attacks. However, there are some problems in the above three aspects of existing reputation-based trust model in online transcation, such as the selection of trust factors.This paper presents a trader reputation-based trust model for online transaction. In this model, a trader has two kinds of reputations system, namely buyer's reputations and seller's reputations, and eah reputation system has two kinds reputation, namely direct reputation and recommend reputation. The direct reputation of a trader relative to another trader is calculated in terms of the reference trader's rating of the transaction between the two traders, whereas the recommend reputation is computed based on all trader s'rating of the transaction between them. To compute the direct and recommend reputations precisely and to obtain stronger resistibility to attacks as well, many comprehensive factors in computing trust value are introduced in the model. To estimate the validity of the rating given by trader s, a quality model and a computational method are also employed to evaluate the objectivity and the credibility of the rating, respectively. To compute the trust value of a user, the concept of belief factor is introduced to integrate the direct reputation with the recommend reputation. Furthermore, a method is put forward in this paper for determining belief factor.Second, based on the presented model, The Reputation Evaluation and Trust Management System is designed, The paper discusses in detail the system System Structure, function structure, data structures and presents the core algorithm.Finally, the effectiveness and resistibility of the proposed trust model are analyzed theoretically and evaluated experimentally. The experimental results show that the proposed trust model outperforms existing models, and can effectively be applied to the online transaction.