In this paper we propose two distinct ways of augmenting the existing clustering environment so that granular data (patterns) can be accommodated. The two approaches deal with either holistic or atomistic representations of metrics and descriptors when mining outcomes of granule-valued random...