Access Type

Open Access Embargo

Date of Award

January 2015

Degree Type


Degree Name



Computer Science

First Advisor

Chandan K. Reddy


Retention of students at colleges and universities has long been a concern for educators for many decades. The consequences of student attrition are significant for both students, academic staffs and the overall institution. Thus, increasing student retention is a long term goal of any academic institution. The most vulnerable students at all institutions of higher education are the freshman students, who are at the highest risk of dropping out at the beginning of their study. Consequently, the early identification of at-risk students is a crucial task that needs to be addressed precisely. In this thesis, we develop a framework for early prediction of student success using survival analysis approach. We propose time-dependent Cox (TD-Cox), which is based on the Cox proportional hazard regression model and also captures time-varying factors to address the challenge of predicting dropout students as well as the semester that the dropout will occur, to enable proactive interventions. This is critical in student retention problem because not only correctly classifying whether student is going to dropout is important but also when this is going to happen is crucial to investigate. We evaluate our method on real student data collected at Wayne State University. The results show that the proposed Cox-based framework can predict the student dropout and the semester of dropout with high accuracy and precision compared to the other alternative state-of-the-art methods.