Decision Tree Modeling for Predicting Research Productivity of University Faculty Members

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.