Abstract

The increasing number of scientific publications also increases the complexity of scientific collaboration, requiring a deeper understanding of the interaction patterns between authors of scientific publications. This study aims to identify collaboration patterns between authors using the Hierarchical Clustering Algorithm (HCA) method based on Jaccard Distance. The evaluation of the method's performance was carried out by comparing the results of HCA-Jaccard with HCA using Euclidean Distance, and the Louvain method as a comparison based on modularity. The results show a Cophenetic Correlation Coefficient (CCC) value of 0.96 for HCA-Jaccard, higher than HCA-Euclidean which is 0.93, indicating that HCA-Jaccard is more consistent. Meanwhile, the Louvain method produces a modularity value of 0.7689 which indicates a fairly good community structure. The novelty of this study is in the application of Jaccard Distance in HCA for scientific collaboration analysis, as well as the integration of centrality metrics to identify the role of authors in more depth in author collaboration. This approach provides insights that can be used by institutions in data-based decision making

Concepts :

Access to Document

Citations by Year