中文摘要：

在图书情报领域，不同的知识单元（如文献、作者、机构、期刊、学科、主题词和关键词等）因各式各样的关联而形成了复杂的知识网络。传统的知识网络构建过程中往往仅依赖知识单元之间的单一关系，因而形成了多个独立的单关系网络，然而真实的知识单元之间往往是多种关系复杂地交织在一起。为解决复杂的知识网络构建这一问题，本文以科研人员这一知识单元为例，提出基于多关系融合的异构社会网络构建方法。将社会网络的构建与社会网络三大核心理论之一的社会资本理论相结合，依据社会资本的三个维度来进行关联测度和融合，生成异构社会网络。针对科研组织的实证研究揭示出异构社会网络能够更好地融合研究人员之间与科研兴趣有关的多种关联，为更精确的社区划分提供帮助。

英文摘要：

In the field of library and information science, different knowledge units （such as document, author, organization, journal, discipline, subject heading list, keywords, are associated in different ways, which lead to complex knowledge network. Traditional approaches to construct knowledge networks rely on single relation between knowledge units. Therefore many isolated single relation networks are constructed. However there are a large number of interconnected multiple relations exist between knowledge units. In order to solve the problem of constructing complex knowledge network, this paper proposed a new way to build researcher network by integrating multiple relations between researchers. In this paper, social capital theory, one of the three core theory of social networks, was adopted in the construction of researcher networks. Multiple relations were extracted according to the three dimensions of social capital, and merged in a heterogeneous social network. An empirical study in research organizations revealed that heterogeneous social networks could better integrate a variety of relationships associated with the people＇s research interest, which also facilitated the accurate community detection.