Feature request: step_mca() or multiple correspondence analysis -- dimension reduction for categorical variables

[exsell-jc]

Similar to step_pca() or step_umap() where it may be useful for dimensional reduction, step_mca() may seem useful for reducing dimensions for categorical predictors.

[juliasilge]

Thanks for this idea @exsell-jc! Do you typically use MASS::mca() or something else for this kind of task?

[exsell-jc]

Thanks for this idea @exsell-jc! Do you typically use MASS::mca() or something else for this kind of task?

Sure thing @juliasilge

Actually, I'm unsure if MASS is better or FactoMineR. I can't really seem to find any benchmarks or comparisons. MASS does seem to be downloaded twice as often and more actively updated when looking at these two links:

https://www.rdocumentation.org/packages/FactoMineR/versions/2.4
https://www.rdocumentation.org/packages/MASS/versions/7.3-56