topic Re: code to run correspondence analysis on m*n contingency matrix as input in Statistical Procedures

code to run correspondence analysis on m*n contingency matrix as input

_Piyush — Tue, 27 Mar 2018 15:35:21 GMT

Hi Team,

I am trying to run correspondence analysis on a m*n contingency matrix, which I am using as input data. I want the plot in 2 dimension. What is the code for it? I tried simple corresponding analysis code, but that requires categorical variables as input data.

Can you please help. Have attached the sample data

Re: code to run correspondence analysis on m*n contingency matrix as input

Reeza — Tue, 27 Mar 2018 15:57:57 GMT

If your data is continuous wouldn't you be using Principal components analysis or factor analysis instead?

@_Piyush wrote:

Hi Team,

I am trying to run correspondence analysis on a m*n contingency matrix, which I am using as input data. I want the plot in 2 dimension. What is the code for it? I tried simple corresponding analysis code, but that requires categorical variables as input data.

Can you please help. Have attached the sample data

Re: code to run correspondence analysis on m*n contingency matrix as input

Anjanita — Tue, 27 Mar 2018 17:07:01 GMT

I guess you want to see it Brand wise. So, please add Brand in the first column. Then use the code below:

PROC CORRESP DATA=XXX outc=corres_result noprint;

var Var1 - Var40 ;

id BRAND;

run;

ods html file="/home/YYY/corresp/xxx_coords_pct.xls";

proc print data=corres_result;

run;

ods html close;

endsas;

Please edit the code as per your dataset & variable name..

Hope this will help.

Re: code to run correspondence analysis on m*n contingency matrix as input

_Piyush — Tue, 27 Mar 2018 18:22:29 GMT

Thanks Anjanita!