Abstract

One of the major techniques for data analysis is Clustering in data mining . In this paper, a partitioning clustering method called the Enhanced K Strange Points Clustering algorithm (EKSPA) is used with Bat algorithm. The Enhanced K Strange points clustering algorithm works by first selecting a point that is the minimum (first strange point) of the dataset. It next selects a point that is furthermost (second strange point) from the minimum and continues till it finds K (as many as the number of clusters) strange points which are farthest and equally spaced from each other. The EKSPA then allots remaining points into clusters closest to these K strange points. Finally, it uses the bat algorithm to select the best (bat) point which may replace the K Strange points as the global best solution (bat) if certain conditions are satisfied or retain the K Strange points as the global best solutions (bats) around which the closest points can cluster. As it has been proven that the Enhanced K Strange points clustering algorithm is computationally faster than the K means clustering algorithm while maintaining the quality of clustering, it is concluded that its combination with the bat algorithm also yields better results than the K Means Bat Algorithm.

Authors and Affiliations