0
$\begingroup$

I do not think I need to log-transform my data as the distribution of the components are not skewed. However, the scale of the components are different (i.e. Age, resting blood pressure, maximum heart rate etc), should I normalise my data before carrying out PCA, Sammon and SOM mapping. I am trying to find if the data is linear and whether it is a hard classification problem.

$\endgroup$