Toba, Hapnes and Wijaya, Evelyn A. and Wijanto, Maresha Caroline and Karnalim, Oscar (2017) Enhanced Unsupervised Person Name Disambiguation to Support Alumni Tracer Study. Global Journal of Engineering Education, 19 (1). pp. 42-48. ISSN 1328-3154
|
Text
4.pdf - Published Version Download (525Kb) | Preview |
Abstract
An alumni database is a valuable information source for the development of a university. However, alumni databases tend to be incomplete. It is always possible for phone numbers and home or e-mail addresses to change. In this study, the authors propose an information collection strategy by gathering information spread across the Internet through search engines. The research is focused on the evaluation of efficiency factors during the name disambiguation process. The authors suggest a combination of reduction (Red-UPND) and supervised queries strategies, which improve the efficiency of the disambiguation process to around 67% compared to the baseline unsupervised person name disambiguation (UPND). The experiment results show that the approach fits the case university’s requirements to support an alumni tracer study and to find people automatically, especially for people with ordinary names.
| Item Type: | Article |
|---|---|
| Uncontrolled Keywords: | Text mining, information extraction, clustering, person name disambiguation, virtual alumni tracer |
| Subjects: | T Technology > T Technology (General) |
| Depositing User: | Perpustakaan Maranatha |
| Date Deposited: | 20 Dec 2017 06:42 |
| Last Modified: | 10 Oct 2018 08:49 |
| URI: | http://repository.maranatha.edu/id/eprint/23793 |
Actions (login required)
 |
View Item |
