Institution Publication Feature Analysis Based on Time-Series Clustering

• Published in: Entropy (Basel, Switzerland) Vol. 24; no. 7; p. 950
• Main Authors: Lin, Weibin, Jin, Mengwen, Ou, Feng, Wang, Zhengwei, Wan, Xiaoji, Li, Hailin
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 07.07.2022; MDPI
• Subjects: Algorithms; associated network; Bibliometrics; Clustering; Cooperation; Data mining; feature analysis; institution publication; Methods; numerical values; Propagation; Time series; time series clustering; trend analysis; China
• ISSN: 1099-4300; 1099-4300
• DOI: 10.3390/e24070950

Based on the time series of articles obtained from the literature, we propose three analysis methods to deeply examine the characteristics of these articles. This method can be used to analyze the construction and development of various disciplines in institutions, and to explore the features of the publications in important periodicals in the disciplines. By defining the concepts and methods relevant to research and discipline innovation, we propose three methods for analyzing the characteristics of agency publications: numerical distribution, trend, and correlation network analyses. The time series of the issuance of articles in 30 important journals in the field of management sciences were taken, and the new analysis methods were used to discover some valuable results. The results showed that by using the proposed methods to analyze the characteristics of institution publications, not only did we find similar levels of discipline development or similar trends in institutions, achieving a more reasonable division of the academic levels, but we also determined the preferences of the journals selected by the institutions, which provides a reference for subject construction and development.