Log-Linear Model Object

DESCRIPTION:

Class of objects that result from fitting log-linear models.

GENERATION:

This class of objects is returned as the paramIter component in the class "missmodel" object returned by the mdLoglin, completeLoglin, emLoglin, and daLoglin functions.

METHODS:

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

STRUCTURE:

The class "Loglin" 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 "Loglin" object, the parameters are the cell probabilities in the log-linear model. The number of parameters equals the number of cells in the table specified by the log-linear model. Using the control argument of Loglin, completeLoglin, daLoglin, or emLoglin, 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:

dimensions
the number of levels of each variable in the table.
dimensionNames
a list of names for each of the levels for each variable in the table.

SEE ALSO:

, , , .