Application of Bayesian Networks in Consumer Service Industry and Healthcare
Le Zhang, Yuan Gao, Balmatee Bidassie, Vincent G. Duffy · Lecture notes in computer science · 2014
Bayesian networks are powerful in data mining and analyzing causal relationships of an uncertain-reasoning problem. The implementation of Bayesian networks in industry and healthcare diagnosis can facilitate the process of locating causations in complex issues. This study conducted two case studies by BayesiaLab in consumer service and healthcare domain. Case Study One used unsupervised learning and supervised learning on the individual data set of county road traffic volume in Indiana State and concluded that road type has the most significant impact on daily vehicle miles traveled. In Case Study Two, only supervised learning was used to observe the aggregated data of adverse mental health effect on civilians, deployed veterans and nondeployed veterans of different genders. Both types of veterans showed higher probability to have adverse mental health compared to civilians. In conclusion, Bayesian networks provided valid results to support prior research. Further research is needed to investigate the differences between using individual data and aggregated data, and to apply Bayesian networks in meta-analysis.