知识演化分析的技术方法研究综述

A Review on Technical Methods for Knowledge Evolution Analysis

[目的/意义]随着前沿科学的快速发展,如何把握相关知识的演化过程成为一项重要且有挑战性的工作。本文旨在梳理知识演化分析的各种技术方法,探析现有技术路线的问题及面临的挑战,为进一步开展相关研究提供思路。[方法/过程]明确知识演化分析的内涵和理论基础,借助文献计量和内容分析的方法,通过VOSviewer对国内外相关研究展开主题分析,在此基础上构建演化分析技术方法图谱。[结果/结论]从知识结构的角度划分出引文网络主路径法、引文网络聚类法、共词网络路径挖掘法、共词网络聚类法、演化机制模型法、主题模型法、深度学习法等技术方法,并归纳出不同技术方法的特点和不足。融合细粒度知识、丰富的知识关联和语义分析技术,提出一种大数据驱动的基于研究设计指纹的知识演化分析框架。

[Purpose/Significance]With the rapid development of frontier science,how to grasp the evolution of relevant knowledge has become an important and challenging task.This paper aims to summarize various technical methods of knowledge evolution analysis,analyze the problems and challenges of existing technical routes,and provide ideas for further relevant research.[Method/Process]The connotation and theoretical basis of knowledge evolution analysis was clarified.With bibliometric and content analysis methods,VOSviewer was used to carry out thematic analysis of related research at home and abroad.Based on it,a method map of evolutionary analysis techniques was constructed.[Result/Conclusion]From the perspective of knowledge structure,7 technical methods are classified into citation network main path analysis,citation network clustering,co-word network path mining,co-word network clustering,evolution mechanism model,topic model,and deep learning method.And summarize the characteristics and shortcomings of different technical methods.By integrating fine-grained knowledge,rich knowledge relationships and semantic analysis technology,a big data-driven knowledge evolution analysis framework based on research design fingerprints is proposed.