基于本体的高校专家定位系统研究

An Ontology-based Experts Locator System within Academia

专家及其拥有的知识是高校最重要的资产和核心竞争力,专家定位系统能够方便人们迅速查找所需领域的专家信息。然而简单的专家列表不能满足用户选择专家的需求。本文的研究目的在于两个方面：一是如何向用户提供丰富的专家信息,以利于用户对专家的比较和选择;二是如何提高专家排序的准确度。我们利用本体来集成体现专家专长的多源异构数据,同时针对专家集成信息中不同的文档类型和结构设定不同的权重,并利用数据融合技术来提高专家排序的有效性。基于这些关键技术,我们构建了专家查询原型系统,并以武汉大学信息管理学院的教师为实验对象进行了初步测试。结果表明,本文所提出的专家查询方法能够获得较高的查准率。

Experts and their knowledge are the most valuable resources and key competencies in higher education institutions.Experts locator system can facilitate users to find suitable experts.However,users may have difficulties in selecting experts only based on a simple list of names.The aim of this paper is two fold：①to provide rich information of experts to users for comparison and selection;②to improve the accuracy of experts ranking results.We propose ontology based integration of multiple heterogeneous data sources.Furthermore we apply field-based weighting models to different document types and structures and use data fusion techniques to improve the ranking efficiency.We build a prototype system and test its efficiency within the School of Information Management,Wuhan University.The results indicate that our method improves the ranking of candidates.