A clustering system for data sequence partitioning

摘要

In data analyzing, data is often presented as sequences. To partition the data sequences, we propose a sequence clustering system in which a fuzzy compatible relation is employed to show the similarity between any two sequences. Moreover, the max–min transitive closure is applied to transfer the fuzzy compatible relation into a fuzzy equivalence relation. It is found that the data sequences with more similar variations are clustered together by using the proposed clustering system. In that case, the sequences are partitioned easily and quickly into clusters.