Abstract
This paper concerns the geometric treatment of graphical models using Bayes linear methods. We introduce Bayes linear separation as a second order generalised conditional independence relation, and Bayes linear graphical models are constructed using this property. A system of interpretive and diagnostic shadings are given, which summarise the analysis over the associated moral graph. Principles of local computation are outlined for the graphical models, and an algorithm for implementing such computation over the junction tree is described. The approach is illustrated with two examples. The first concerns sales forecasting using a multivariate dynamic linear model. The second concerns inference for the error variance matrices of the model for sales, and illustrates the generality of our geometric approach by treating the matrices directly as random objects. The examples are implemented using a freely available set of object-oriented programming tools for Bayes linear local computation and graphical diagnostic display.
Similar content being viewed by others
References
Dawid A.P. 1979. Conditional independence in statistical theory. J. Roy. Statist. Soc. B 41(1): 1–31.
Dawid A.P. and Lauritzen S.L. 1993. Hyper Markov laws in the statistical analysis of decomposable graphical models. Ann. Statist. 21: 1272–1317.
de Finetti B. 1974. Theory of Probability, Vol. 1. Wiley.
Goldstein M. 1981. Revising previsions: a geometric interpretation. J.R. Statist. Soc. B 43: 105–130.
Goldstein M. 1988a. Adjusting belief structures. J.R. Statist. Soc. B 50: 133–154.
Goldstein M. 1988b. The data trajectory. In Bernardo J.-M. et al. (Eds.), Bayesian Statistics 3. Oxford University Press, pp. 189–209.
Goldstein M. 1990. Influence and belief adjustment. In Smith J. and Oliver R. (Eds.), Influence Diagrams, Belief Nets and Decision Analysis. Wiley, Chichester.
Goldstein M. 1997. Prior inferences for posterior judgements. In Chiara M.L.D. et al. (Eds.), Structures and Norms in Science. Kluwer, Pordrecht.
Goldstein M. 1999. Bayes linear analysis. In Encyclopedia of Statistical Sciences. Wiley, Chichester. Update volume 3.
Goldstein M., Farrow M., and Spiropoulos T. 1993. Prediction under the influence: Bayes linear influence diagrams for prediction in a large brewery. The Statistician 42(2): 445–459.
Goldstein M. and Wooff D.A. 1995. Bayes linear computation: concepts, implementation and programming environment. Statistics and Computing 5: 327–341.
Jensen F.V. 1996. An Introduction to Bayesian Networks. UCL Press.
Lauritzen S.L. 1992. Propagation of probabilities, means, and variances in mixed graphical association models. J. Amer. Statist. Assoc. 87, 420: 1098–1108.
Lauritzen S.L. 1996. Graphical Models. Oxford Science Publications.
Lauritzen S.L., Dawid A.P., Larsen B.N., and Leimer H.G. 1990. Independence properties of directed Markov fields. Networks 20: 491–505.
Normand S.-L. and Tritchler D. 1992. Parameter updating in a Bayes network. J. Amer. Statist. Assoc. 87, 420: 1109–1115.
Pearl J. 1988. Probabilistic Reasoning in Intelligent Systems. Morgan Kaufmann.
Smith J.Q. 1990. Statistical principles on graphs. In Smith J. and Oliver R. (Eds.), Influence Diagrams, Belief Nets and Decision Analysis. Wiley, Chichester.
Tierney L. 1990. LISP-STAT: An Object Oriented Environment for Statistical Computing and Dynamic Graphics. Wiley.
Wilkinson D.J. 1997. BAYES-LIN: An object-oriented environment for Bayes linear local computation. U. Ncle. Stats. Preprint STA97, 20.
Wilkinson D.J. 1998. An object-oriented approach to local computation in Bayes linear belief networks. In Payne R. and Green P. (Eds.), Proceedings in Computational Statistics 1998. Physica-Verlag, Heidelberg, pp. 491–496.
Wilkinson D.J. and Goldstein M. 1996. Bayes linear adjustment for variance matrices. In Bernardo J.-M. et al. (Eds.), Bayesian Statistics 5. University Press, Oxford, pp. 791–800.
Wilkinson D.J. and Goldstein M. 1997. Bayes linear covariance matrix adjustment for multivariate dynamic linear models. U. Ncle. Stats. Preprint STA97, 12.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Goldstein, M., Wilkinson, D.J. Bayes linear analysis for graphical models: The geometric approach to local computation and interpretive graphics. Statistics and Computing 10, 311–324 (2000). https://doi.org/10.1023/A:1008977409172
Issue Date:
DOI: https://doi.org/10.1023/A:1008977409172