Improvement of k-Means Clustering Algorithm by GA

Khamis H Haroun, Wu Zhifeng

Abstract: As known that K-means algorithm is the one of the common and popular technique for solving clustering problems. In fact, there are many k-means algorithms for solving the clustering problem such as Lloyds k-means clustering algorithm, hierarchical k-means algorithm, also Grid based k-means algorithm etc. In the classical k-means algorithm the selected value of k must be confirmed first. So, the resulting clusters mainly depends on the selection of the initial centroids. It is not simple job to select the accurately value of k or to know exactly number of clusters for the given data set. So that in this paper propose new algorithm that called improvement of k-means clustering algorithm by GA that algorithm will be able to automatic find the best initial centers and appropriation of clusters according to the given data set.

Keywords: Clustering, K-means, Cluster centroid, Genetic algorithm

How to Cite?: Khamis H Haroun, Wu Zhifeng, "Improvement of k-Means Clustering Algorithm by GA", Volume 7 Issue 4, April 2018, International Journal of Science and Research (IJSR), Pages: 1429-1435, https://www.ijsr.net/getabstract.php?paperid=ART20181981, DOI: https://dx.doi.org/10.21275/ART20181981

Download Citation: APA | MLA | BibTeX | EndNote | RefMan