Correspondence Analysis

A medical application of correspondence analysis


Table of contents

En francais s'il vous plait


1. Introduction
  1. Data Analysis
  2. Factor Analysis

2. Overview of the Method

  1. A little bit of History of Linguistics

3. The Analysis I

  1. The Indenpendency Hypothesis
  2. Khi 2 and Inertia
  3. Profiles

4. The Analysis II

  1. Dimensionnality of the Problema
  2. Geometrical Principles
  3. Factorial Axis
  4. Symetry of both Analysises

5. What you Find in the Output and How to Use it

  1. Absolute distance, relative distance
  2. Eigenvalues and Inertia of the Table
  3. How many Axis to Keep ?
  4. Factorial Scores
  5. Absolute Contributions by Point
  6. Absolute Contributions by Cell
  7. Relative Contributions by Point and by Axis

6. Graphical Representations and Interpretation

  1. The assymetric graph
  2. The symetric graph
  3. Interpretation of Distances Between Points of the Same Cloud
  4. Interpretation of Distances Between Points of Diiferent Cloud
  5. Angular Interpretation Between a Point and an Axis
  6. Two Typical Patterns
  7. Interpretation of my First axis
  8. Interpretation of my Second Axis

7. Extensions and Limitations

  1. Supplementary Points
  2. Multiple Correspondence Analysis
  3. Extensions and Exemples
  4. Softwares

8. Bibliography


This paper was first published at Ramses Abul Naga's Advanced Econometrics Workshop at the HEC in Lausanne in 1996. The same year, I recycled it for Francois Bavaud Multivariate Statistics course at the SSP in the University of Lausanne. The introduction was written with an all-economists audience in mind, but this does not reduce the generality of the text neither its usefulness for other classes of readers.

François-Xavier Micheloud, Pully, November 1997

My Homepage