Please use this identifier to cite or link to this item: http://dspace.utpl.edu.ec/handle/123456789/18782
Title: A Bayesian network approach to classifying bad debt in hospitals
Authors: Shi, D.
Keywords: Bad debt recovery
Bayesian network
Classification
Healthcare industry
metadata.dc.date.available: 2017-06-16T22:02:21Z
Issue Date: 7-Mar-2016
Publisher: Proceedings of the Annual Hawaii International Conference on System Sciences
Abstract: The rising bad debts for unpaid medical treatments in hospitals pose serious problems in many countries. Researchers have started to use computational intelligence methods to construct models to classify bad debt as an important first step in debt recovery. However, the academic research dealing with this issue has been scarce. Previous studies have examined bad debt situations where only a small number of independent attributes were available, thus leaving out many potentially relevant factors in bad debt recovery. In this study, we used a richer data set containing bad debt cases from a hospital. The objective of the study was to explore the effectiveness of using a Bayesian network to classify the bad debt through comparison with alternative methods in different scenarios. The results show that the Bayesian network-based models have the best classification accuracy rates and exhibit the best global performance at most probability cutoffs and significantly outperform other models. The conditional probability distribution generated by the Bayesian network models reveals the important attributes and their relationships. The results can help hospitals identify the related characteristics of patient-debtors, look for better potential solutions, and better manage medical bad debt. © 2016 IEEE.
metadata.dc.identifier.other: 10.1109/HICSS.2016.412
URI: http://dspace.utpl.edu.ec/handle/123456789/18782
ISBN: 15301605
ISSN: 978-076955670-3
Other Identifiers: 10.1109/HICSS.2016.412
Other Identifiers: 10.1109/HICSS.2016.412
metadata.dc.language: Inglés
metadata.dc.type: Article
Appears in Collections:Artículos de revistas Científicas

Files in This Item:
There are no files associated with this item.


This item is protected by original copyright



Los recursos publicados en el RiUTPL se distribuyen bajo la licencia: CreativeCommons