Le Hoai Minh, Ta Minh Thuy, Le Thi Hoai An, Pham Dinh Tao: "DC Programming and DCA for clustering using weighted dissimilarity measures".

Abstract: The purpose of this paper is to develop new efficient approachbased on DC (Difference of Convex functions) programming and DCA (DC Algorithm) for clustering using weighted dissimilarity measures which is formulated as a hard combinatorial optimization problem. DC reformulation technique and exact penalty in DC programming are developed to build an appropriate equivalent DC program which leads to an explicit DCA scheme. Numerical results on real word datasets show the efficiency, the scalability of DCA and its superiority with respect to standard algorithm.

Keywords:


Citation: Le Hoai Minh, Ta Minh Thuy, Le Thi Hoai An and Pham Dinh Tao. DC Programming and DCA for clustering using weighted dissimilarity measures, Proceedings of the 5th NIPS Workshop on Optimization for Machine Learning, Lake Tahoe, Nevada, USA, December 2012, 4 pages.

 

Download link