Hierarchical clustering for multiple nominal data streams with evolving behaviour

Abstract Over the decade, a number of attempts have been made towards data stream clustering, but most works fall under clustering by example approach. There are applications where variable approach is required which involves multiple streams as opposed to examples in stream. Furthermore, few presented for and these applicable numeric only. Hence, this research gap has motivated current work. In present work, hierarchical technique proposed cluster nominal. To address concept changes splitting merging clusters structure performed. The decision split or merge based on entropy measure, representing cluster’s degree disparity. performance analysed compared Agglomerative Nesting synthetic well real-world dataset terms Dunn Index, Modified Hubert $$\varGamma $$ Γ statistic, Cophenetic Correlation Coefficient, Purity. outperforms evolving streams. effect evolution average visualised detailed analysis understanding.