MatrixSim: A new method for detecting the evolution paths of research topics

• Published in: Journal of informetrics Vol. 16; no. 4; p. 101343
• Main Authors: Wang, Xiaoguang, He, Jing, Huang, Han, Wang, Hongyu
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.11.2022
• Subjects: Co-word network; Evolution path detection; Matrix similarity; MatrixSim; Research topic evolution; Research topic evolution; Matrix similarity; Co-word network; MatrixSim; Evolution path detection
• ISSN: 1751-1577; 1875-5879
• DOI: 10.1016/j.joi.2022.101343

•A new method for detecting the evolution paths of research topics called MatrixSim is proposed.•The community structure of co-word networks is maximally preserved.•The similarity of topic communities in terms of both words and relations is integrated through MatrixSim.•Two sets of experiments are designed to verify the validity and feasibility in research topic evolution analysis.•Methodological support is provided for researchers to conduct prospective research activities. In this study, MatrixSim, a new method for detecting the evolution paths of research topics based on matrix similarity, was proposed. In the analysis of research topic evolution with the help of co-word networks, in contrast to traditional methods of topic evolution path detection, such as cosine similarity and edge similarity, MatrixSim is based on the local community structure of topic communities in co-word networks and considers the similarity of research topics in both nodes and edges, that is, words and inter-word relations. Using the library and information science field as an example, two sets of experiments were designed for topic similarity detection and subject-specific research topic evolution analysis to evaluate and verify the performance of MatrixSim in detecting the evolution paths of research topics and its validity and feasibility in research topic evolution analysis. The results confirm that MatrixSim performs well in detecting the evolution paths of research topics. It can correlate important research topics, help describe the research development process in scientific fields, reveal the internal evolutionary features of research topics, and thus discover and track the research frontiers in scientific fields. This study provides significant methodological support for researchers conducting prospective research activities.