基于文献知识抽取的专题知识库构建研究——以中药活血化瘀专题知识库为例

Research on Construction of a Subject Knowledge Base based on Literature Knowledge Extraction: Using the Knowledge Base of Activating Blood Circulation and Removing Stasis as the Object

大数据时代科研人员对高效获取和利用领域知识提出了更高的要求,文献作为科研人员快速准确地了解本领域研究状况的有效途径,基于文献的知识发掘已成为一种新的科研方式。专题知识库作为组织和管理某一特定领域知识的工具,能够用于挖掘和展现文献背后的知识以满足用户个性化需求。本文提出了面向特定研究问题的专题知识库建设路线,采用基于知识工程的信息抽取方法,通过抽象研究问题要素构建专题知识模型,将其作为信息抽取的知识模式,制定知识模型各节点的知识抽取策略,对文献中实体、关系及属性进行解析、抽取与关联组织,基于这些结构化知识提供知识检索、浏览、问答、可视化关联组织等一系列知识服务。然后以中药活血化瘀领域建设实践为例,详细阐述了基于文献知识抽取构建专题知识库的实施方案。系统功能测试显示,该专题知识库能够实现知识快速查询、知识与文献关联发现、知识结构梳理等预期服务场景。本研究提供了一种构建专题知识库行之有效的技术路线,能够帮助科研用户快速而准确地定位和获取文献中的深层知识,提供了数据密集型科研环境下学科化资源建设与个性化精准服务的转型方式。

Researchers put forward higher requirements for efficient acquisition and utilization of domain knowledge in the big data era. As literature is an effective way for researchers to quickly and accurately understand the research situation in their field, knowledge discovery based on literature has become a new research method. As a tool to organize and man‐age knowledge in a specific domain, the subject knowledge base can be used to mine and present the knowledge behind the literature to meet users’ personalized needs. This paper designs the construction route of the subject knowledge base for specific research problems. An information extraction method based on knowledge engineering is adopted. First, a subject knowledge model is built through abstraction of the research elements. Then, under the guidance of the knowledge model,the knowledge extraction strategy of each model node is developed to analyze, extract, and correlate entities, relations, and attributes in the literature. Finally, a database platform based on this structured knowledge is developed that can provide a variety of services such as knowledge retrieval, knowledge browsing, knowledge Q&A, and visualization correlation. Tak‐ing construction practices in the field of activating blood circulation and removing stasis as an example, this paper analyzes how to construct a subject knowledge base based on literature knowledge extraction. As the system functional test shows,this subject knowledge base can realize the expected service scenarios such as quick query of knowledge, related discovery of knowledge and literature, and knowledge organization. As this study proposes an effective technical route to building a subject knowledge base to help researchers locate and acquire deep knowledge in literature quickly and accurately, it pro‐vides a transformation mode of resource construction and personalized precision services in the data-intensive research en‐vironment.