Nonsmooth DC programming approach to the minimum sum-of-squares clustering problems

作者：

Highlights：

• Characterizations of stationary points of the clustering problems are given.

• An algorithm for solving the clustering problem using DC representation is developed.

• An algorithm for solving the clustering problem based on the DCA method is developed.

• The proposed algorithms are tested using large and very large data sets.

摘要

Highlights•Characterizations of stationary points of the clustering problems are given.•An algorithm for solving the clustering problem using DC representation is developed.•An algorithm for solving the clustering problem based on the DCA method is developed.•The proposed algorithms are tested using large and very large data sets.

论文关键词：Cluster analysis,Non-smooth optimization,Non-convex optimization,Incremental clustering algorithms

论文评审过程：Received 2 April 2015, Revised 16 October 2015, Accepted 15 November 2015, Available online 28 November 2015, Version of Record 8 February 2016.

论文官网地址：https://doi.org/10.1016/j.patcog.2015.11.011