9

In an answer to this question about treating categorical data as continuous, optimal scaling was mentioned. How does this method work and how is it applied?

Community
  • 1
Freya Harrison
  • 3,212
  • 4
  • 25
  • 31

1 Answers1

7

The general idea is that you should scale the categorical variable in such way that the resulting continuous variables will be just the most useful. So, it is always coupled with some regression or learning procedure and so the fitting of the model is accompanied by optimization (or trying various possibilities) of ordinal variables scaling.
For some more practical issues, consult the docks of R aspect and homals packages.