Dear All,
I wondered if anyone was aware of a predictive clustering procedure in R for mixed data types?
For example, suppose that I had conducted an unsupervised k-prototypes procedure in order to estimate clusters for health states.
Is it possible to take the clusters from the k-prototypes procedure, and use a predictive/supervised procedure to estimate clusters at time t+1, t+2 etc?
Would appreciate all feedback.
Thank you