DocumentCode :
3762310
Title :
Decision tree modeling for predicting research productivity of university faculty members
Author :
Arfika Nurhudatiana;Adilla Anggraeni
Author_Institution :
School of Information Systems, Bina Nusantara University, JWC Campus, Jalan Hang Lekir 1 No. 6, Senayan, Jakarta, Indonesia
fYear :
2015
Firstpage :
70
Lastpage :
75
Abstract :
Research productivity is a common issue for university management, especially in most teaching-based universities. This research aims to investigate the various factors contributing to the research productivity of university faculty members through the use of decision tree modeling. Three separate models, each for junior, intermediate, and senior faculty members, were developed to quantitatively predict individual faculty member´s publication output in a specific target year. The models were trained and evaluated on a dataset containing academic profiles and past publications of 78 full-time faculty members in Binus University International by using cross-validation. The overall accuracy of the models was above 80%, with the model for the junior faculty members achieved 100% accuracy. Several key findings were identified from this research. First, possession of doctoral degree was found to be a key identifier of productive junior faculty members. Second, length of service with university did not affect research productivity. Third, tenured faculty members were identified to be productive in research after two years of consistent publications.
Keywords :
"Productivity","Decision trees","Medical services","Data mining","Predictive models","Data models","Education"
Publisher :
ieee
Conference_Titel :
Data and Software Engineering (ICoDSE), 2015 International Conference on
Print_ISBN :
978-1-4673-8428-5
Type :
conf
DOI :
10.1109/ICODSE.2015.7436974
Filename :
7436974
Link To Document :
بازگشت