Last modified: 2018-05-18
Abstract
Additive Bayesian networks (ABNs) are types of graphical models that extend the usual generalized linear model (GLM) to multiple dependent variables through the representation of joint probability distribution. Thanks to their flexible properties, ABNs have been widely used in epidemiological analyses. In this work we present a veterinary case study where ABNs are used to explore multivariate swine diseases data of medical relevance. We then compare the results with a classical methodology. Finally, we highlight the key difference between a multivariable standard (GLM) and a multivariate (ABN) approach: the latter attempts not only to identify statistically associated variables, but to additionally separate these into those directly and indirectly dependent with one or more outcome variables
References
1. Pittavino, M.: Additive Bayesian Networks for Multivariate Data: Parameter Learning, Model Fitting and Applications in Veterinary Epidemiology. PhD thesis. Universitaet Zuerich (2016)
2. Pittavino, M., Dreyfus, A., Heuer, C., Benschop, J., Wilson, P., Torgerson, P., Furrer, R.: Comparison between Generalized Linear Modelling and Additive Bayesian Network. Identification of Factors associated with the Incidence of Antibodies against Leptospira interrogans sv Pomona in MeatWorkers in New Zealand. Acta Tropica 173, 191-199, ELSEVIER (2017)
3. Sanchez-Vazquez, M.J., Nielen, M., Edwards, S.A., Gunn, G.J. and Lewis F.I.: Identifying associations between pig pathologies using a multidimensional machine learning methodology. BMC Vet. Res. 8:151, 1-11 (2012)
4. Rijmen, F.: Bayesian networks with a logistic regression model for the conditional probabilities, Int. Jour. of Appr. Reas. 48:2, 659-666 (2008)
5. R Development Core Team. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org (2017)
6. Kratzer, G., Pittavino, M., Lewis, F., Furrer, R.: abn: an R package for modelling multivariate data using additive Bayesian networks. The Comprehensive R Archive Network, 1-37 (2017)
7. Friedman, N., Goldszmidt, M., Wyner, A.: Data analysis with Bayesian networks: A Bootstrap approach, Proc. 15th Conf. on Uncert. in Artif. Intell. (UAI’99), San Francisco: Morgan Kaufmann, 196-205 (1999)
8. Koivisto, M., Sood, K. Exact Bayesian structure discovery in Bayesian networks. Jour. of Mach. Lear. Res. 5, 549-573 (2004)
9. Plummer, M.: JAGS: a program for analysis of Bayesian graphical models using Gibbs Sampling. Proc. 3rd Int. Work. Dist. Stat. Comp. (DSC 2003), Vienna, Austria, 1-10 (2003)
10. Pittavino, M., Dreyfus, A., Heuer, C., Benschop, J.,Wilson, P., Torgerson, P., Furrer, R.: Data on Leptospira interrogans sv Pomona in Meat Workers in New Zealand. Data in Brief 13, 587-596, ELSEVIER (2017)