Asymmetric clustering using the alpha–beta divergence

作者：

Highlights：

• We propose the use of asymmetric dissimilarity in centroid-based clustering.

• We asymmetrize the AB-divergence on the basis of the within-cluster variances.

• The proposed approach efficiently takes into consideration clusters with different variances.

• Experiments were conducted on real and simulated data of high and low dimensionality.

摘要

Highlights•We propose the use of asymmetric dissimilarity in centroid-based clustering.•We asymmetrize the AB-divergence on the basis of the within-cluster variances.•The proposed approach efficiently takes into consideration clusters with different variances.•Experiments were conducted on real and simulated data of high and low dimensionality.

论文关键词：Clustering,Asymmetry,Dissimilarity,Alpha–Beta divergence

论文评审过程：Received 10 October 2012, Revised 16 November 2013, Accepted 20 November 2013, Available online 1 December 2013.

论文官网地址：https://doi.org/10.1016/j.patcog.2013.11.019