Data discovery using rough set based reductive partitioning: some experiments

An experimental investigation into unsupervised database mining is conducted. A novel paradigm for autonomous mining based on recursive partitioning is tested. The speculation is that increasing coherence will increase conceptual information. This in turn will reveal previously unrecognized, useful and implicit information. To assist our partitioning heuristics, a rough set based methodology called Total Roughness is designed to measure the crispness of a partition. This methodology is used in our experiments to help in partitioning as well as perform non-scalar data clustering. What is particularly noteworthy is that our approach sometimes partitions on multiple attributes. The feasibility of integrating rough set theory in unsupervised partitioning is evaluated and addressed. We focus on the use of rough sets. We also provide some discussion of our other experiments