Canonical Analysis - Notes and Technical Information

Confirmatory models. If you have specific hypotheses about the relationships between the different variables in your data, you can test those hypotheses via the Structural Equation Modeling (SEPATH) module. With SEPATH, you can fit (confirmatory) factor models, path models, models with intercepts, etc. Also, models can be fit to multiple samples, and parameter estimates can be constrained (to be equal) across samples.

The general linear model. The General Linear Models (GLM) and General Regression Models (GRM) modules are full implementations of the general linear model, and will perform all of the same types of analyses available in canonical correlation.

Partial Least Squares Regression (PLS). The General Partial Least Squares (PLS) module is particularly well suited for analyzing problems with very many predictor variables and one or more dependent variables.