基于细粒度关键词引用网络的领域知识多维分析

Multi-dimensional Analysis of Domain Knowledge Based on a Fine-grained Keyword Citation Network

针对关键词引用网络中存在的节点语义角色缺失和关联关系单一的局限,本文通过学术文本语义功能增强网络中节点及其关联关系的语义信息,提出了一种细粒度关键词引用网络。首先对学术文本内容进行解析,抽取文献关键词、引用关联、引文上下文、引用对象等信息,并对其词汇功能和引用功能进行识别。接着采用复杂网络图方法构建细粒度关键词引用网络,从引用功能敏感的子网分析、特定节点的多维关联分析和细粒度领域知识演化分析三个方面进行领域知识多维分析,并以ACL(Association for Computational Linguistics)会议论文集为例进行实证研究。结果验证了本文所提出方法的有效性,发现了NLP(natural language processing)领域知识间的使用、扩展和对比模式,揭示了特定研究问题的发展情况或特定方法的应用情况,刻画了领域知识的细粒度演化脉络。本文扩展了知识网络的研究方法和深度,也为领域知识多维分析提供了新的视角和路径。

Due to the lack of a semantic role of nodes and the association relationship between nodes being single in the current keyword citation network,this study enhances the semantic information of nodes and their association relationship through an academic text semantic function and proposes a fine-grained keyword citation network.First,this paper analyzes the content of academic text,extracts the keywords,citation associations,citation context,and citation objects of academic literature and identifies their term and citation functions.Next,the complex network method is utilized to construct a fine-grained keyword citation network.The multi-dimensional analysis of domain knowledge is carried out based on three aspects:citation function-aware subnet analysis,multi-dimensional association analysis of specific nodes,and fine grained domain knowledge evolution analysis.The ACL domain dataset is taken as an example for empirical research.Results verify the validity of the proposed method;discover the patterns of use,extension,and comparison among domain knowledge;reveal the development of specific research problems or the application of specific methods;and describe the fine-grained evolution of domain knowledge. This study extends the research method and research depth of keyword networks and provides a new perspective and path for multi-dimensional analysis of domain knowledge.