post: Posterior probabilities/intercepts, and mass point...

In npmlreg: Nonparametric Maximum Likelihood Estimation for Random Effect Models

Description

Takes an object of class glmmNPML or glmmGQ and displays the posterior probabilites w_ik as well as the posterior intercepts (Sofroniou et. al, 2006). Further it classfies the observations to mass points according to their posterior probability. The level on which the information in all three cases is displayed can be chosen by the user via the level argument ("upper" or "lower"). The actual information in both cases is identical, the latter is just an expanded version of the former. In case of simple overdispersion models, the level argument is not relevant.

Usage

post(object, level="upper")

Arguments

object	an object of class glmmNPML or glmmGQ.
level	"upper" or "lower".

Value

A list of the following four items:

prob	posterior probabilities (identical to object$post.prob in case of "lower" and for one-level models).
int	posterior intercepts (identical to object$post.int in case of "lower" and for one-level models).
classif	a numerical vector containing the class numbers (the order of the classes corresponds to the order of the mass points given in the output of alldist or allvc).
level	either "lower", "upper", or "none" (for one-level models).

Author(s)

Jochen Einbeck and John Hinde (2006)

References

Sofroniou, N., Einbeck, J., and Hinde, J. (2006). Analyzing Irish suicide rates with mixture models. Proceedings of the 21st International Workshop on Statistical Modelling in Galway, Ireland, 2006.

See Also

alldist, allvc

Examples

 data(galaxies, package="MASS")
 gal <- as.data.frame(galaxies)
 post(alldist(galaxies/1000~1, random=~1, data=gal, k=5))$classif
    # classifies the 82 galaxies to one of the five mass points
 

npmlreg documentation built on May 2, 2019, 9:31 a.m.