中文摘要：

对等网信誉系统的一个关键问题是如何提高推荐信息的可用性。现有模型将推荐节点的交易可信度等同于其推荐可信度，因而容易造成恶意推荐节点对信誉系统的虚假推荐和共谋推荐攻击。提出了一种基于意图隐藏的推荐可信度评价模型。在该模型中，一方面恶意推荐节点无法判断节点的查询意图，因而很难采取针对性行为；另一方面，基于历史推荐满意度的评价方法能对节点的推荐可信度进行有效评价。分析和仿真结果验证了模型的有效性。

英文摘要：

A key issue for peer-to-peer network reputation systems is how to improve the usability recommenda- tory information. In the existing models, the trustworthiness of transaction for the recommender is viewed as equal as that of recommendation, as a result, the reputation systems oan be easily attacked by the malicious recommender through the false or collusive recommendation. A model for evaluating the- trustworthiness of recommendation based on obscure intention is proposed. In the model, as the malicious recommender could not identify the query intention of the nodes, it＇s difficult for the malicious to response. On the other hand, the evaluating method based on the satisfaction of historic recommendation can efficiently evaluate the trustworthiness of the recommendation. The analysis and simulation result prove its efficiency.