Data Clustering for Optimized Information Search with Hybrid Evolutionary Approaches
Anuradha D. Thakare
Manuscript received on 06 February 2019 | Revised Manuscript received on 19 February 2019 | Manuscript Published on 04 March 2019 | PP: 291-294 | Volume-7 Issue-5S2 January 2019 | Retrieval Number: ES205001751/19©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Clustering is an important data analysis technique which reveals the relationships among unexplored data objects. Cluster initialization and selection of seeds in first iteration contributes to the quality of clustering. The prime objective is to find best cluster with some quality measure. K-means is prone to local optima since initial centroids are selected randomly. In order to evaluate this problem, some heuristic clustering algorithms are introduced along with evolutionary approaches like Genetic Algorithms and Swarm Intelligence. Genetic Algorithms are the heuristic search techniques and are found to be robust to envisage the optimal or near optimal combination of weights in a multidimensional space. This article presents comparative analysis of various hybrid evolutionary approaches developed for clustering to find the optimal cluster center. The objective is to improve the quality of clusters. From the analytical and experimental results, it is observed that the proposed hybrid evolutionary algorithms perform satisfactorily over the existing approaches. As compared to hybrid PSOBA, Multi Stage Genetic Clustering results into reduced error rate by 30 to 50 percent for thyroid and iris dataset respectively. The clustering results vary with respect to dataset and the internal spread.
Keywords: Clustering; Evolutionary Algorithms; Genetic Algorithms(GA); Particle Swarm Optimization(PSO); Bee Algorithm(BA); K- means(KM).
Scope of the Article: Clustering