Functional data clustering

We propose a novel global-local framework that partitions coefficients into exclusive subsets with independent Dirichlet process (DP) priors on each subset to improve the scalability, while simultaneously providing enhanced flexibility compared to global clustering in the high-dimensional function.