PCA pre-treatment: Centering or Standardizing?

[TiagoDias]

Hello all,

I have a matrix X that contains industrial data, where each column is a variable.

Since PCA is done by the correlation matrix, should I use:
. original X;
. X after Centering (subtracting each column of X by the mean of the respective column);
.X after Standardizing? (subtracting each column of X by the mean and dividing by the standard deviation);

I think autoscaling is better, but I dont know if it is a must or just a good practice

Thanks for your time

 

[GretaGarbo]

.X after autoscaling? (subtracting each column of X by the mean and dividing by the standard deviation);

It is called standardizing.

("autoscaling" is a phrase from a strange software. The subject should define the software, not the other way around.)

 

[TiagoDias]

It is called standardizing.

It's not from a software, in some papers it is called auto-scaling, but ok. i update my comment to standardizing.
("autoscaling" is a phrase from a strange software. The subject should define the software, not the other way around.)