On Predictive Distributions for the Generalized Growth Curve Model

by Jack C. Lee, J. W. Lee and Y. L. Hsu .

Abstract: Due to the complexity of predictive distributions for the generalized growth curve model, we propose some useful approximations. Using the Gibbs sampling, we can compare, via a real data set, the simulations with the approximate predictive distributions to obtain the best approximation among the candidates considered.

Key Words: Approximations, Gibbs sampling, Metropolis-Hastings, Laplacian, Rao-Blackwellization, Ramus height data

Jack C. Lee, jclee@stat.nctu.edu.tw
J. W. Lee, jwlee@stat.nctu.edu.tw
Y. L. Hsu, ylhsu@stat.nctu.edu.tw

Editor: Dayanand N. Naik , dnn@math.odu.edu

READING THE ARTICLE: You can read the article in portable document (.pdf) format (256495 bytes.)

NOTE: The content of this article is the intellectual property of the authors, who retains all rights to future publication.

This page has been accessed 2453 times since July 24, 2006.

Return to the InterStat Home Page.