The HPFMM Procedure

Parameterization of Model Effects

PROC HPFMM constructs a finite mixture model according to the specifications in the CLASS, MODEL, and PROBMODEL statements. Each effect in the MODEL statement generates one or more columns in the matrix $\bX$ for that model. The same $\bX$ matrix applies to all components that are associated with the MODEL statement. Each effect in the PROBMODEL statement generates one or more columns in the matrix $\bZ$ from which the linear predictors in the model for the mixture probability models is formed. The same $\bZ$ matrix applies to all components.

The formation of effects from continuous and classification variables in the HPFMM procedure follows the same general rules and techniques as for other linear modeling procedures. For information about constructing the model effects, see the section Specification and Parameterization of Model Effects of Chapter 4: Shared Statistical Concepts.