It would be useful to have tools for Classification for nominal variables. Specifically, for cluster tools for binary variables/ mixed variables (i.e., some nominal and Continous variables at the same time). This would also be used for CATA data. Possible solutions could include- latent class, binary clustering algorithms, with binary data might only consider 1's or 0s for clustering (currently use Latent Gold and Sawtooth, CCEA software). For perspective, we like that these other programs run k-means multiple times and look across all runs to see which is most reproducible, allows nominal and continuous, provides optimal number of clusters, and provides fit measurements (goodness of fit) to help determine optimal number of clusters.
