eMixMargDen(bayesm)R Documentation

Compute Marginal Densities of A Normal Mixture Averaged over MCMC Draws

Description

eMixMargDen assumes that a multivariate mixture of normals has been fitted via MCMC (using rnmixGibbs). For each MCMC draw, the marginal densities for each component in the multivariate mixture are computed on a user-supplied grid and then averaged over draws.

Usage

eMixMargDen(grid, probdraw, compdraw)

Arguments

grid array of grid points, grid[,i] are ordinates for ith component
probdraw array - each row of which contains a draw of probabilities of mixture comp
compdraw list of lists of draws of mixture comp moments

Details

length(compdraw) is number of MCMC draws.
compdraw[[i]] is a list draws of mu and inv Chol root for each of mixture components.
compdraw[[i]][[j]] is jth component. compdraw[[i]][[j]]$mu is mean vector; compdraw[[i]][[j]]$rooti is the UL decomp of Sigma^-1.

Value

an array of the same dimension as grid with density values.

Warning

This routine is a utility routine that does not check the input arguments for proper dimensions and type. To avoid errors, call with output from rnmixGibbs.

Author(s)

Peter Rossi, Graduate School of Business, University of Chicago, Peter.Rossi@ChicagoGsb.edu.

References

For further discussion, see Bayesian Statistics and Marketing by Allenby, McCulloch, and Rossi.
http://gsbwww.uchicago.edu/fac/peter.rossi/research/bsm.html

See Also

rnmixGibbs


[Package bayesm version 0.0 Index]