Abstract Paper

Journal of Computing and Intelligent Systems - JCIS

Title	: AN OPTIMIZED CLUSTER CENTER INITIALIZATION USING K-MEANS AND CLUSTERING WITH LARGE APPLICATIONS
Author(s)	: N. Nivetha, R. Pugazendi
Article Information	: Volume 2 - Issue 2 (December - 2018) , 45-49
Affiliation(s)	: Research Scholar, Dept. of Computer Science, Government Arts College, Periyar University, Salem – 7, Tamilnadu, India.
	: Assistant Professor, Dept. of Computer Science, Government Arts College, Periyar University, Salem – 7, Tamilnadu, India.

Abstract :

Cluster analysis is an essential and unsupervised data mining technique that is used for grouping objects automatically. Partitional clustering algorithms gather an individual partition of the data rather than a clustering structure. K-means is the most popular, simple and efficient method. The computational complexity of k-means cannot have problems with the size of the data set. One of the major drawbacks of Traditional k-means clustering algorithm’s is the selection of initial centroids and do not work with huge datasets. To solve this problem a new hybrid technique has been proposed for k-means clustering and Clara. This algorithm is more efficient for working with large datasets. We have estimated the performance using three real data sets. The proposed technique is much more effective for improving accuracy and reduces computation time as well as iterations.

Keywords	: Data mining, Clustering, K-Means, Clara, Euclidean distance, Complexity
Document Type	: Research Paper
Publication date	: November 27, 2018