Data Mining A PeopleSoft Database To Assist In Developing Student Retention Interventions

Abstract

Per a Bellwether Education Partners study (Aldeman, 2015, p. 8), "As of 2013, there were 29.1 million college dropouts versus 24.5 million Americans who dropped out with less than a high school diploma. In pure, raw numbers, college dropouts are a bigger problem than high school dropouts." Conceptually this study is framed within theories of student persistence/attainment and the Knowledge Discovery Process (KDP). This research study developed first time in college (FTIC) and transfer (TRAN) student graduation prediction models by using decision trees and support vector machine (SVM) classification algorithms and identified attributes of students who graduate and do not graduate. Data was collected from the University of Houston’s data warehouse to provide detailed student academic records as the basis for quantitative analysis. The data set included male and female undergraduate students enrolled in the College of Education’s Teaching & Learning Program from 2000-2012 at the University of Houston. These findings may contribute to improving student success and subsequent graduation rates in the College of Education and other colleges across the campus.

Description

Keywords

Knowledge discovery process (KDP), Data mining, PeopleSoft, Student retention, Databases, Student intervention, Persistence, Attainment

Citation