Improving Bregman k-means. (1st January 2014)

• Record Type: Journal Article
• Authors: Ashour, Wesam; Fyfe, Colin
• Is Part Of: International journal of data mining, modelling and management. Volume 6:Number 1(2014)
• Page Start: 65
• Page End: 82
• Publication Date: 2014-01-01
• Subjects: clustering -- K-means -- local optima -- Bregman divergences; Data mining -- Periodicals; Information science -- Periodicals; Databases -- Periodicals; 005.7
• Languages: English
• ISSNs: 1759-1163

Abstract

We review Bregman divergences and use them in clustering algorithms which we have previously developed to overcome one of the difficulties of the standard k-means algorithm which is its sensitivity to initial conditions which leads to finding sub-optimal local minima. We show empirical results on artificial and real datasets.