1995/06/05 by Darren J Wilkinson, Michael Goldstein, Wilkinson, Darren J +1
Physics and Astronomy · #Data Analysis #FOS: Physical sciences #Statistics and Probability (physics.data-an) #bayes-an #physics.data-an
paper · pdf · doi:10.48550/arxiv.bayes-an/9506002
In submission. LaTeX, 17 pages, Chicago BIB-style (included). Also available as a postscript file from http://fourier.dur.ac.uk:8000/~dma3djw/djwgdlm.html For information about [B/D], go to http://fourier.dur.ac.uk:8000/stats/bd/
arxiv created 1995/06/05 · arxiv updated 2009/12/01
A methodology is developed for the adjustment of the covariance matrices underlying a multivariate constant time series dynamic linear model. The covariance matrices are embedded in a distribution-free inner-product space of matrix objects which facilitates such adjustment. This approach helps to make the analysis simple, tractable and robust. To illustrate the methods, a simple model is developed for a time series representing sales of certain brands of a product from a cash-and-carry depot. The covariance structure underlying the model is revised, and the benefits of this revision on first order inferences are then examined.