Design and Implementation of a Recommender System for use at an Emergency Homeless Shelter in Calgary
Date
2021-05-04
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Modern homeless shelters are collecting data from key interactions with clients. This data can be utilized by machine learning algorithms to identify clients that are at risk for chronic homelessness. This would provide shelter operators with a powerful new tool to assist them in housing individuals. However, most machine learning algorithms are not suitable for the task due to the lack of interpretability. Classification rule learning is brought forward in this work as an exceedingly interpretable class of machine learning algorithms. A novel recommender system based on classification rule learning is proposed and evaluated on local homeless shelter data. The results from this work suggest that classification rule learning is robust and interpretable enough to be used to support modern homeless shelters.
Description
Keywords
Homeless Shelter, Recommender System, Identification, Rule Learning, Classification
Citation
John, C. T. (2021). Design and Implementation of a Recommender System for use at an Emergency Homeless Shelter in Calgary (Master's thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca.