Off-campus Eastern Washington University users: To download EWU Only theses, please use the following link to log into our proxy server with your EWU NetID and password.

Non-EWU users: Please talk to your local librarian about requesting this thesis through Interlibrary loan.

Date of Award

Spring 2021


Access perpetually restricted to EWU users with an active EWU NetID

Document Type

Thesis: EWU Only

Degree Name

Master of Science (MS) in Mathematics




Homelessness remains an issue in the United States in spite of increased efforts to assist those experiencing it. In recent years there has been a growing focus on prevention programs that aim to provide assistance to those who would, if assistance were not provided, experience homelessness. A key part of these homelessness prevention programs is identifying who to assist. Statistical models can play the important role of predicting who is at highest risk of experiencing homelessness. In this study, customer utility billing data, as well as data related to housing and geographical area, are analyzed and tested for usefulness in the prediction of homelessness. It was found that the utility billing variables were not significantly associated with the outcome of homelessness; however, Logistic Regression, Cox Proportional Hazard, and the Cox Time Varying Covariates models found these variables to be useful in predicting homelessness. These models were evaluated for prediction using the method of K-Folds and the Cox Time Varying Covariates model was found to have the best performance with a maximum precision of 33% while predicting only 1/302 people experiencing homelessness or, using a different prediction binning threshold, a maximum f-1 Score of 8% with a precision of 5% and a recall of 8%. Model evaluation varies depending on outcome binning threshold, which can be adjusted to suit different types of homelessness prediction programs. This performance is essentially the same as other homelessness prediction models in the literature. Model performance would likely improve with additional data more related to homelessness such as income, social service usage, and medical service usage.