Correspondence Analysis

Correspondence analysis plays a role similar to factor analysis or principal component analysis for categorical data expressed as a contingency table (e.g. as described in the chi-square test of independence).

Essentially, correspondence analysis decomposes the chi-square statistic of independence into orthogonal factors. This approach is valid even when the cell sizes in the contingency table are less than 5 (or even zero).

Topics

References

Rencher, A.C., Christensen, W. F. (2012) Methods of multivariate analysis (3nd Ed). Wiley

Johnson, R. A. and Wichern, D. W. (2007) Applied multivariate statistical analysis. 6th Ed. Pearson.
https://www.webpages.uidaho.edu/~stevel/519/Applied%20Multivariate%20Statistical%20Analysis%20by%20Johnson%20and%20Wichern.pdf

 

 

12 thoughts on “Correspondence Analysis”

  1. Hi Charles,
    You are doing a phenomenal work as I have been reading through your content and replies on website for my project. Thank you for such a quality content on Regression and Statistics on Excel. I want to perform Logistics Regression on employee attrition dataset where Y is Attrition Status of the employee (Yes/No) and X are Age, Department, Job Role, Monthly Income, Marital Status, etc. Overall I have 44 independent variables like these (30 before one-hot encoding). I have been trying to find ways to implement FAMD (Factor Analysis of Mixed Data) to decrease the number of independent variables before performing the regression. I am particularly looking for FAMD for this purpose because my independent variables are of mixed type i.e., 14/30 features are numerical features, say, Age, Monthly Income, Distance From Home, Percent Salary Hike, etc., while 16/30 features are categorical, say, Department (HR, Sales, R&D), Gender (M/F), Job Role (Sales Representative, Sales Executive, Manager, Laboratory Technician, etc.), Job Level (1,2,3,4), and so on. Is there a way to implement FAMD using Real Statistics Add-in in Excel? And how to interpret the data derived from running the FAMD command, if any, as well as how to use the reduced dimensions for Logistic Regression model implementation?

    Reply
  2. Doctor Zainontz, buenos días, muchas gracias, por tan grande beneficio que nos presta a la comunidad de investigadores, con su página. Implementaría Ud. el Análisis de correspondencia múltiple, en una versión siguiente? y el DOE 3^k?

    Dr. Zainontz, good morning, thank you very much, for the great benefit you give us to the research community, with your page. Would you implement the Multiple Correspondence Analysis in a subsequent version? and the DOE 3 ^ k?

    Reply

Leave a Comment