Access Type

Open Access Thesis

Date of Award

January 2013

Degree Type


Degree Name



Computer Science

First Advisor

Chandan Reddy


Hospital readmissions are not only expensive but are also potentially harmful, and most importantly, they are often preventable. Providing special care for a targeted group of patients who are at a high risk of readmission can significantly improve the chances of avoiding rehospitalization. Despite the significance of this problem, not many researchers have thoroughly investigated it due to the inherent complexities involved in analyzing and estimating the inherent predictive power of such complex hospitalization records. In this thesis, we propose using support vector machines and survival analysis methods to analyze data collected from Electronic Medical Records (EMR). We define the notion of abnormal patients and understand how they affect the performance of classifiers. We use sparse methods with survival regression models to build clinical models which are suitable to apply on such complex clinical data. These models are compared with existing readmission models such as ADHERE, TABAK and logistic regression models. Finally, we provide inferences and conclusions on how to extend this work to build better regression models.