Conditional Multivariate Gaussian Model Object

DESCRIPTION:

Class of objects that result from fitting conditional multivariate gaussian models.

GENERATION:

This class of objects is returned as the paramIter component in the class "missmodel" object returned by the mdCgm, completeCgm, emCgm, and daCgm functions.

METHODS:

The class "cgm" has associated methods: VarNames.cgm , , , , and .

STRUCTURE:

The class "cgm" object consists of a matrix. Each row in the matrix represents one set of parameters estimates, typically produced using an iterative algorithm like EM or data augmentation. In the "cgm" object, the parameters are the cell probabilities, cell means and the lower triangle of the variance-covariance matrix of a multivariate normal (Gaussian) distribution. Using the control argument of mdCgm, completeCgm, daCgm, or emCgm, you can choose which iterates to save. The iteration number for each row of saved estimates is given by the row name.

SPECIAL ATTRIBUTES:

VALUE:

dimensionNames
a list of names of the levels for each variable in the table.
varnames
list of names of the variables in the original data, in the same order as in the original data.
xbar
vector of means subtracted from the original continuous variables when standardizing the variables.
sdv
vector of scale factors. The original variables are standardized by subtracting the means, and dividing the result by the appropriate scale.

SEE ALSO:

, , , .