Please use this identifier to cite or link to this item:
http://hdl.handle.net/1942/42787
Title: | A new approach to computing the distances between research disciplines based on researcher collaborations and similarity measurement techniques | Authors: | VANCRAEYNEST, Bram PHAM, Hoàng Son ALI ELDIN, Amr |
Issue Date: | 2024 | Publisher: | Source: | Journal of Informetrics, 18 (3) (Art N° 101527) | Abstract: | The measurement of distance between research disciplines involves various approaches, with a focus on publication citation analysis. However, calculating discipline distance requires more than just selecting relevant information; it also involves choosing suitable quantification methods and similarity measures. In this paper, we introduce a novel approach to measuring the distance between research disciplines, referred to as a distance matrix. This approach is particularly useful when there is limited availability of citation data, providing an alternative method for quantifying the distance between disciplines. Our method counts co-occurrences of disciplines based on researcher collaborations in projects and evaluates various similarity measures to convert the co-occurrence matrix into a similarity matrix. We analyze the behavior of different similarity measures and propose functions to transform the similarity matrix into a distance matrix, capturing research discipline dissimilarity effectively. Additionally, we establish evaluation criteria for distance matrix quality. We implement our approach on the Flanders Research Information Space dataset, showing promising results. The distance matrix demonstrates satisfactory density scores, outperforming traditional approaches in skewness and deviation. The probability density functions of distances remain consistent over time, indicating stability. Furthermore, the distance matrix proves valuable for visualizing discipline profiles associated with the dataset, providing valuable insights. | Keywords: | Distance metrics;Research collaboration;Research disciplines;Similarity measures | Document URI: | http://hdl.handle.net/1942/42787 | ISSN: | 1751-1577 | e-ISSN: | 1875-5879 | DOI: | 10.1016/j.joi.2024.101527 | Rights: | 2024 Elsevier Ltd. All rights reserved. | Category: | A1 | Type: | Journal Contribution |
Appears in Collections: | Research publications |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
publication.pdf Restricted Access | Published version | 1.91 MB | Adobe PDF | View/Open Request a copy |
Article_A_new_approach_to_computing_the_disciplines.pdf Until 2024-11-20 | Peer-reviewed author version | 2.15 MB | Adobe PDF | View/Open Request a copy |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.