Authors : Yugal Kumar and G. Sahoo
Abstract: K-Means algorithm is most popular partition based algorithm that is widely used in data clustering. A lot of algorithms have been proposed for data clustering using K-Means algorithm due to its simplicity, efficiency and ease convergence. In spite this K-Means algorithm has some drawbacks like initial cluster centers. In this study, a new method is proposed to address the initial cluster centers problem in K-means based on binary search technique. The initial cluster centers is obtained using Binary Search Method and the newly generated cluster centers are used as initial cluster centers in K-means to gain optimal cluster centers in dataset. The performance of the Proposed algorithm is tested on the two benchmark dataset iris and wine that are downloaded from the UCI machine learning repository and compare the proposed method with Random, Hartigan and Wang, Ward, Build, Astrhan, Minkowaski ward and IWKM Method in which proposed method with K-means provides 82.93 and 68.94 accuracy rate and intra cluster distance is 105.72 and 18059.81 with iris and wine datasets as well as proposed method with IWKM provides 96.7 and 95.8 accuracy rate.
Yugal Kumar and G. Sahoo, 2014. An Initialization Method for K-Means Algorithm Using Binary Search Technique. International Journal of Soft Computing, 9: 131-137.