Transactions on Machine Learning and Data MiningP-ISSN: 1865-6781, E-ISSN 2509-9337ISBN 978-3-942952-72-9
Volume 12 - Number 2 - October 2019 - Pages 33-54
Applications of Bayesian Networks
Ron Kenett
KPA Ltd., Raanana and Samuel Neaman Institute, Technion, Israel
Abstract
Modelling relationships between variables has been a major challenge for statisti-cians in a wide range of application areas. This capability is an essential component in an effort to generate information of high quality from a given data set. Bayesian Networks (BN) combine graphical analysis with Bayesian analysis to represent relations linking measured and target variables. Such graphical maps can be used for diagnostics and predictive analytics. The paper presents applica-tions of Bayesian Networks to various domains such the evaluation of web site usability, the testing of web services, operational risks, biotechnology, customer satisfaction surveys, healthcare systems and an analysis of the impact of management style on statistical efficiency. These case studies provide complementary aspects of BN applications to emphasize the breath of potential applications and the various associated methodological challenges. Following the presentation of these case studies, a general section discusses various properties of Bayesian Networks, including the study of causality with BNs. Some references to soft-ware programs used to construct BNs are also provided. A concluding section summarises the paper main points and lists current research topics.
Keywords: Bayesian Networks, Directed Acyclic Graph, Conditional Probability Distribu-tion, Information Quality.
Download Paper
(379 KB)