|
|
Scholars' Label Expansion Method Combining Topic Similarity and Co-authorship Network
|
Sheng Jiaqi,Xu Xin
|
(Department of Information Management, Faculty of Economics and Management, East China Normal University, Shanghai 200062, China)
|
|
|
Abstract
[Objective] In order to predict the future research direction and interest of scholars, this paper designs the method of extracting and expanding academic labels from scholar abstracts. [Methods] The basic academic labels are extracted from the abstract through the TF-IDF method. Combined with the topic similarity between scholars and the cooperative relationship between scholars, the basic academic labels are expanded using the labels of similar scholars and other scholars in the team. [Results] Compared with directly using scholars’ current academic labels to predict scholars' future academic labels, using labels expanded by combining topic similarity and co-authorship network increases recall rate by 8.33% on average. [Limitations] The sample size is small, the method is only for single-language papers and cannot cover other language papers published by scholars. The universality of method still needs further confirmation. [Conclusions] The method proposed in this paper has certain predictive power for scholars' future research directions and research interests.
|
Published: 21 May 2020
|
|
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
|
Shared |
|
|
|
|
|
Discussed |
|
|
|
|