基于细粒度知识图谱的科技文献主题发现与热点分析

Topic Discovery and Hotspot Analysis of Scientific Literature Based on Fine-Gained Knowledge Graph

[目的/意义]利用知识图谱进行细粒度的知识组织并识别科技文献的热点主题,有助于科研工作者把握领域的研究现状和学科前沿,进而为科研资源优化配置提供有力支持。[方法/过程]提出一套基于知识图谱的主题发现和热点分析方法。首先,识别科技文献中的知识元,再抽取知识元中的语义实体进行主题发现;其次,构建“文献—知识元—主题”知识图谱,通过知识图谱中的共现关系网络识别领域的研究热点;最后,以农学领域的中文学术论文为实验数据,进行实证研究。[结果/结论]所构建的细粒度知识图谱不仅能够揭示科学知识与主题的潜在关联,而且能够实现科技文献主题的热点分析。

[Purpose/significance]Utilizing knowledge graphs for fine-grained knowledge organization and identifying hot topics in scientific literature is beneficial for researchers to grasp the current status and frontiers of their field.This approach can provide robust support for optimizing the allocation of research resources.[Method/process]Propose a knowledge graph-based method for topic discovery and hotspot analysis.Firstly,identify knowledge elements in scientific literature and extract semantic entities from these elements for topic discovery.Secondly,construct a“literature-knowledge element-topic”knowledge graph and identify research hotspots in the field through co-occurrence relationship networks within the knowledge graph.Finally,empirical research was conducted using Chinese academic papers in the field of agriculture as experimental data.[Result/conclusion]The research results indicate that the fine-grained knowledge graph constructed in this study not only reveals the potential associations between scientific knowledge and topics but also accomplishes hotspot analysis of themes in scientific literature.