中文摘要：

开放百科是互联网最重要的参考信息源,吸引了大规模贡献者的参与,然而实践社区缺乏对用户贡献行为可信度的系统自动评估。本文借鉴可信计算的理念,提出了一条基于间接反馈的评估路径,采用分句粒度的文本分析法,以开放百科内容编辑史为数据源展开计量分析,构造了用户间反馈值、剩余贡献比等量化评估指标,并逐步设计了算法流程,以实现对用户贡献行为可信度的系统自动评估。Wikipedia开源数据的实证分析进一步表明该评估路径和方法具有可行性,评估结果具有实践意义。

英文摘要：

Open encyclopedia, the most important reference source on Internet, has attracted large- scale participation of contributors; however, it lacks systematically automatic assessment on credibility of user contribution behaviors. Using trust computing for reference, the paper put forward an assessment approach based on indirect feedback, which used clausal text analysis method and took editing history records as data source to design some quantifiable assessment indexes such as feedback values between users and the ratio of remainder to original. It designed an algorithm process step by step to accomplish systematically automatic assessment on credibility of user contribution behaviors. The empirical analysis based on Wikipedia open source data indi- cates the feasibility of our assessment approach and the practical significance of the assessment results.