Applications of Bayesian Networks
Chapman & Hall/CRC series in computer science & data analysis/Series in computer science and data analysis · 2010
Modelling relationships between variables has been a major challenge for statisticians 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 applications 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 software programs used to construct BNs are also provided. A concluding section summarises the paper main points and lists current research topics.