A multidisciplinary ensemble algorithm for clustering heterogeneous datasets

Clustering is a commonly used method for exploring and analysing data where the primary objective to categorise observations into similar clusters. In recent decades, several algorithms methods have been developed clustered data. We notice that most of these techniques deterministically define cluster based on value attributes, distance, density homogenous single-featured datasets. However, definitions are not successful in adding clear semantic meaning clusters produced. Evolutionary operators statistical multidisciplinary may help generating meaningful Based this premise, we propose new evolutionary clustering algorithm (ECA*) social class ranking meta-heuristic stochastically heterogeneous multifeatured The ECA* integrated with recombinational operators, Levy flight optimisation, some techniques, such as quartiles percentiles, well Euclidean distance K-means algorithm. Experiments conducted evaluate against five conventional approaches: (KM), K-means++ (KM++), expectation maximisation (EM), learning vector quantisation (LVQ), genetic clustering++ (GENCLUST++). That end, 32 datasets examine their performance using internal external basic measures measure how sensitive features (cluster overlap, number clusters, dimensionality, structure, shape) form an operational framework. results indicate surpasses its counterpart terms ability find right Significantly, compared less properties mentioned above. Thus, order overall algorithms, from best performing worst performing, ECA*, EM, KM++, KM, LVQ, GENCLUST++. Meanwhile, rank 1.1 (where 1 represents 6 refers algorithm) dataset