Principal Components & Classification Analysis - General Purpose

Data are often collected on variables that are not only correlated, but also are large in number. This makes the interpretation of the data and the detection of its structure difficult. By transforming the original variables to a smaller number of uncorrelated variables, Principal Components Analysis (PCA) makes these two jobs easier. The Principal Components and Classification Analysis (PCCA) implemented in STATISTICA aims to achieve two objectives:

  1. Reduction in the number of variables to a smaller number of, 'representative' and 'uncorrelated' variables.
  2. Classification of variables and cases.

The methods used in this module are similar to those offered in the Factor Analysis module, but differs in the following ways:

  1. PCCA does not use any iterative methods to extract factors.
  2. PCCA allows you to consider some variables and/or cases as supplementary (refer to Principal Components & Classification Analysis - Data Reduction for a description of supplementary variables and cases). These variables and cases can be mapped onto the same factor space as derived from the analysis (active) variables and cases.
  3. PCCA allows you to analyze the data collected on variables that are heterogeneous with respect to their means or with respect to both their means and standard deviations, by providing an option to analyze covariance matrices as well as correlation matrices.

The sections listed in Principal Components & Classification Analysis - Introductory Overview describe the principles of PCCA. In order to understand these principles you need to know the basic statistical concepts described in Elementary Concepts. Your familiarity with the concepts of variance/covariance and correlation, in particular, can help you a great deal in understanding the principles of PCCA as well as for interpreting the results produced by PCCA. A useful reference for technical details of the PCA implemented in this module is Jambu, 1991. Also, because the methods implemented in PCCA are similar and related to the techniques available in the Factor Analysis module, you may want to review the Overviews for that module as well.