句子级知识抽取在情报学中的应用分析

Analysis of the Application of the Sentence-level Knowledge Extraction in Information Science

通过比较句子级知识抽取与词语级知识抽取的差异性,分析句子级知识抽取在情报学中的意义,表现在四类典型应用系统:学术抄袭检测系统、参考文献自动标注系统、文献自动综述系统、知识库构建系统。分析了知识抽取的难点与关键技术,针对难点与关键技术提出了知识抽取的3个转向:抽取对象转向以学术文献为主;抽取技术转向以内容结构分析为主;抽取目标转向以构建知识元数据库为主。

Based on the comparison of the differences between sentence-level knowledge extraction and word- level knowledge extraction, this paper analyzes the significance of the sentence-level knowledge extraction in information science. It＇s represented in the 4 typical application systems ： academic plagiarism detection system, automatic reference-labeling system, automatic literature summarizing system and knowledge base construction system. After analyzing the difficulties and key technologies of knowledge extraction, this paper proposes 3 shifts of knowledge extraction in accordance with the difficulties and key technologies ： extraction object shifts to giving priority to academic literature; extraction technology shifts to giving priority to content structure analysis; and extraction objective shifts to giving priority to constructing knowledge metadata base.