Multivariate Methods of Data Analysis

Multivariate data matrices are matrices or supermatrices, containing more than two variables, either continuous, binary, or the binary-continuous combinations. The multivariate-continuous data matrix

 

 

 

can be readily associated with methods of factor analysis, or other methods for analysis of structure, as various methods of cluster analysis. The multivariate continuous data supermatrix

 

 

 

is a prototypical data matrix of the linear multiple regression. The multivariate continuous data supermatrix

 

 

 

is a prototype of data supermatrix used in canonical analysis. The multivariate binary-continuous data supermatrix

 

 

 

is a prototype of the various analysis of variance methods, and the multivariate continuous-binary data supermatrix

 

 

 

is a prototype of the multivariate discriminant analysis. Other variations, combining continuous variables with the coding variables as submatrices within the data supermatrix are possible, defining data matrices for such methods as, e.g., analysis of covariance. These variations will be described in the chapters discussing these methods of data analysis.