How to Deal with Nominal and Ordinal Measures?

Learn to deal with nominal and ordinal measures.

Solution for nonparametric data

As discussed earlier, PCA works mostly for ratio or interval measures. So, what happens if our data is categorical (nonparametric, i.e., measured using an ordinal or nominal scale) or mixed? There are different methods to use for different types of data:

  • PCA is used for ratio and interval measures.

  • MCA (Multiple correspondence analysis) is used for categorical data.

  • MFA (Multiple factor analysis) is used for ratio, interval, or categorical data, but the data needs to be of the same type for any given group. There is an R package called MFAmixdata that includes a function called `MFAmix` to perform this type of analysis but with mixed data.

  • PCAMIX is used for a mix of categorical and ratio and interval measures. There is an R package called PCAmixdata that includes functions for handling mixed data as well as PCA and MCA.

Get hands-on with 1200+ tech skills courses.