Electrical and Computer Engineering ETDs

Publication Date

Summer 7-31-2017


A college student's success depends on many factors including pre-university characteristics and university student support services. Student graduation rates are often used as an objective metric to measure institutional effectiveness. This work studies the impact of such factors on graduation rates, with a particular focus on delay in graduation. In this work, we used feature selection methods to identify a subset of the pre-institutional features with the highest discriminative power. In particular, Forward Selection with Linear Regression, Backward Elimination with Linear Regression, and Lasso Regression were applied. The feature sets were selected in a multivariate fashion. High school GPA, ACT scores, student's high school, financial aid received, and first generation status were found to be important for predicting success. In order to predict delay in graduation, we trained predictive models using Support Vector Machines (SVMs), Gaussian Processes (GPs), and Deep Boltzmann Machines (DBMs) on real student data. The difference in performance among the models is negligible with respect to overall accuracies obtained. Further analysis showed that DBMs outperform SVMs in terms of precision and recall for individual classes. However, the DBM and SVM implementations are computationally inexpensive compared to GPs, given the same resources.


Higer Education, Analytics, Machine Learning, Data Mining, Predicitve Analytics, Feature Selection, Graduation Delay, Graduation Rate

Document Type




Degree Name

Computer Engineering

Level of Degree


Department Name

Electrical and Computer Engineering

First Committee Member (Chair)

Gregory L Heileman

Second Committee Member

Manel Martinez-Ramon

Third Committee Member

Don R Hush

Fourth Committee Member

Ahmad Slim