Standard Article

Inverted Wishart Distribution, Generalized

Statistical Theory and Methods

  1. Philip J. Brown

Published Online: 15 SEP 2006

DOI: 10.1002/9780470057339.vag008

Encyclopedia of Environmetrics

Encyclopedia of Environmetrics

How to Cite

Brown, P. J. 2006. Inverted Wishart Distribution, Generalized. Encyclopedia of Environmetrics. 3.

Author Information

  1. University of Kent at Canterbury, UK

Publication History

  1. Published Online: 15 SEP 2006

Abstract

The generalized inverted Wishart (GIW) distribution is a flexible extension of the inverted Wishart (IW) distribution which overcomes the major deficiency of that distribution whilst retaining much of the ease of computation inherent in the natural conjugate roots of the inverted Wishart distribution in Bayesian analysis for normal sampling. In offering a more flexible prior distribution for the covariance matrix of a multivariate normal distribution it has general applicability. It also offers a conjugate Bayesian analysis for problems of missing data where the missing data has a monotone structure. It has been especially useful in simultaneous modeling of the spatial distribution of ozone where monitoring stations come on-stream in sequence, using multivariate regression analysis for interpolation and prediction.