5

The Normal-inverse-Wishart distribution is a conjugate prior for the multivariate normal distribution when the mean and covariance are unknown. I understand that conjugate priors are mathematically convenient but are there specific applications where the justification for using the Normal-inverse-Wishart distribution prior goes beyond convenience?

kjetil b halvorsen
  • 77,844
Mykie
  • 511
  • 1
    I would suggest https://stats.stackexchange.com/questions/155059/justification-for-conjugate-prior for consideration – beuhbbb Aug 08 '17 at 15:28

1 Answers1

2

One application is the Gibbs Sampling from a Dirichlet Process mixture model, where a conjugate prior is required. See page 33 of the pdf below

https://www.cs.cmu.edu/~kbe/dp_tutorial.pdf

George
  • 21
  • 2
  • 1
    Hi @George, welcome to XV. OP seems to be asking about when conjugacy's justification goes beyond its being required or convenience. – Taylor Sep 06 '18 at 21:00