Dear all,
As a follow-up to last week's email regarding standardisation in PCA, I have
now established that covariance-based PCA is preferable when the data are
measured in the same units and are on a Likert (1...5 discrete category
scale). Performing a correlation-based PCA in such a case will inflate the
importance of a variable having little variation. Performing a
covariance-based PCA will allow the 'natural' variation of variables to be
reflected in the PCA.
As an addition, if my variables were in the same units (say length
measurmements of various components) but were measured on a continuous
scale, would correlation or covariance based PCA be preferable? I have seen
various examples in texts which use correlation based PCA in such a case.
Many thanks again,
Kim.
Hello all,
I am going to do a principal components analysis. I know that we can
standardise the variables (correlation-based PCA) to make them 'equally
important'.
I have read that
1)if we omitted standardisation, a variable which varied lot would tend to
dominate the principal components, and
2) standardisation to make variables equally important is suggested when the
variables are measured in different units.
My question is: if my variables are all measured in the same units (say
each variable contained scores for n people and each score could take one of
5 values 1 to 5), would it still be OK to do a correlation based PCA (i.e.
using covariance matrix of standardised variables) or would covariance based
PCA (I.e. using covariance matrix of unstandardised variables) be more
appropriate?
Many thanks, in advance for your help,
All the Best,
Kim.
|