Methods of Dimensionality Reduction in Survival Analysis: An Application in Prediction of Hospital Readmission




Sumner, Joel

Journal Title

Journal ISSN

Volume Title



Hospital Readmissions are a major cost on hospitals with a large source being largely preventable. In order to improve the quality of care and reduce costs, hospitals have employed machine learning methods to analyze electronic health records in order to better predict hospital readmissions. The challenge of this is larger volumes of data take more time to develop and implement prediction models for. In this thesis, we seek to determine if using dimensionality reduction with machine learning methods can improve the performance of machine learning models. The machine learning methods we will use are Artificial Neural Networks (ANN), Cox Hazard Regression Model, Random Forest, K Nearest Neighbor (KNN), Support Vector Machines, Subspace Discriminant, Boosted Trees, and Bagged Trees. The dimensionality reduction methods used were Principal Component Analysis (PCA), Kernel Principal Component Analysis (KPCA), Isomap, Laplacian Eigenmaps, and Locally Preserving Projections (LPP). The performance of the models is evaluated using Area Under the Curve (AUC) and cases where dimensionality reduction improves the performance are reported.


This item is available only to currently enrolled UTSA students, faculty or staff. To download, navigate to Log In in the top right-hand corner of this screen, then select Log in with my UTSA ID.




Mechanical Engineering