Variational learning of a Dirichlet process of generalized Dirichlet distributions for simultaneous clustering and feature selection

作者：

Highlights：

• An infinite generalized Dirichlet mixture for simultaneous clustering and feature selection is proposed.

• A variational approach is developed to learn the proposed model.

• The proposed statistical framework is validated by artificial data as well as two real challenging applications.

摘要

•An infinite generalized Dirichlet mixture for simultaneous clustering and feature selection is proposed.•A variational approach is developed to learn the proposed model.•The proposed statistical framework is validated by artificial data as well as two real challenging applications.

论文关键词：Infinite mixture models,Dirichlet process,Generalized Dirichlet,Feature selection,Clustering,Images categorization,Image auto-annotation

论文评审过程：Received 25 April 2012, Revised 27 February 2013, Accepted 26 March 2013, Available online 17 April 2013.

论文官网地址：https://doi.org/10.1016/j.patcog.2013.03.026