一种面向主题覆盖度与权威度的评审专家推荐模型研究

发布时间：2018-03-02 23:12

本文选题：专家推荐　切入点：主题覆盖度　出处：《图书情报工作》2017年01期 　论文类型：期刊论文

【摘要】：[目的 /意义]为投稿论文遴选出合适的审稿专家是论文发表过程中关键的一环。随着投稿论文和候选评审专家数量的持续增长,人工指定评审专家的方法在准确性和公平性上的弊端日益显露出来。因此,为进一步提高专家评审的客观性和准确性,笔者从专家知识与专家权威度两个维度对专家建模,并以此为依据为不同主题的投稿论文遴选推荐评审专家。[方法 /过程]首先分析专家知识以及投稿论文的研究内容,并提取两者涉及的多个子研究主题;然后,计算专家知识对投稿论文子主题的覆盖度,并提出融合主题特征与时间特征的权威度算法TTAM来分析专家权威度;最后,提出融合主题覆盖度和专家权威度的专家推荐框架CAUFER,综合考虑覆盖度和权威度两个因素为投稿论文推荐合适的评审专家。[结果/结论]实验结果表明,与经典的基于向量空间模型、语言模型和作者主题模型3种专家推荐算法相比,笔者提出的算法能够较好地提高专家与投稿论文的匹配度,并可据此追踪专家权威度的变化,刻画专家在特定主题下的权威度,进一步提高专家推荐的准确性和科学性。
[Abstract]:[objective / significance] the selection of suitable reviewers for contribution papers is a key part of the publication process. As the number of contributing papers and candidate review experts continues to grow, Therefore, in order to further improve the objectivity and accuracy of the expert review, the author models the expert from the two dimensions of expert knowledge and expert authority. On the basis of this, the selection and evaluation experts of contribution papers on different topics are selected. [methods / processes] first analyze the expert knowledge and the research content of the contribution papers, and extract the multiple sub-research topics involved in the two articles; then, To calculate the coverage of expert knowledge to the subtopic of the contribution paper, and propose an authority degree algorithm TTAM to analyze the expert degree of authority, which combines the feature of topic with the feature of time. An expert recommendation framework, CAUFER, which combines topic coverage and expert authority, is proposed, and two factors of coverage and authoritativeness are taken into account to recommend suitable evaluation experts. [results / conclusions] the experimental results show that, Compared with the classical three expert recommendation algorithms based on vector space model, language model and author's subject model, the algorithm proposed by the author can improve the matching degree between experts and contribution papers, and track the change of expert authority. Describe the authority of experts under specific topics, and further improve the accuracy and scientific nature of expert recommendations.
【作者单位】：北京师范大学政府管理学院;
【基金】：中央高校基本科研业务费专项资金资助项目“基于社会网络关系的智能专家遴选与推荐平台建设”(项目编号:SKZZB2014037) 教育部人文社会科学研究青年基金项目“面向论文评审专家推荐的兴趣变化挖掘与回避机制生成的研究”(项目编号:16YJC870006) ISTICEBSCO文献大数据发现服务联合实验室基金项目“融合异构科研数据的评审专家推荐研究”研究成果之一
【分类号】：TP391.3

【参考文献】