getparam.mix {mix}

Present Parameters of General Location Model in an Understandable Format

Description

Present parameters of general location model in an understandable format.

Usage

getparam.mix(s, theta, corr=FALSE)

Arguments

s

summary list of an incomplete normal data matrix created by the function prelim.mix.

theta

list of parameters such as one produced by the function em.mix, da.mix, ecm.mix, or dabipf.mix.

corr

if FALSE, returns a list containing an array of cell probabilities, a matrix of cell means, and a variance-covariance matrix. If TRUE, returns a list containing an array of cell probabilities, a matrix of cell means, a vector of standard deviations, and a correlation matrix.

Value

if corr=FALSE, a list containing the components pi, mu and sigma; if corr=TRUE, a list containing the components pi, mu, sdv, and r.

The components are:

pi

array of cell probabilities whose dimensions correspond to the columns of the categorical part of $x$. The dimension is c(max(x[,1]),max(x[,2]),...,max(x[,p])) where p is the number of categorical variables.

mu

Matrix of cell means. The dimension is c(q,D) where q is the number of continuous variables in x, and D is length(pi). The order of the rows, corresponding to the elements of pi, is the same order we would get by vectorizing pi, as in as.vector(pi); it is the usual lexicographic order used by S and Fortran, with the subscript corresponding to x[,1] varying the fastest, and the subscript corresponding to x[,p] varying the slowest.

sigma

matrix of variances and covariances corresponding to the continuous variables in x.

sdv

vector of standard deviations corresponding to the continuous variables in x.

r

matrix of correlations corresponding to the continuous variables in x.

Note

In a restricted general location model, the matrix of means is required to satisfy t(mu)=A%*%beta for a given design matrix A. To obtain beta, perform a multivariate regression of t(mu) on A — for example, beta <- lsfit(A, t(mu), intercept=FALSE)$coef.

References

Schafer, J. L. (1996) Analysis of Incomplete Multivariate Data. Chapman & Hall, Chapter 9.

See Also

prelim.mix, em.mix, ecm.mix, da.mix, dabipf.mix.

Examples

data(stlouis)
s <- prelim.mix(stlouis,3)    # do preliminary manipulations
thetahat <- em.mix(s)   # compute ML estimate
getparam.mix(s, thetahat, corr=TRUE)$r   # look at estimated correlations