Simultaneous feature selection and Gaussian mixture model estimation for supervised classification problems

作者：

Highlights：

• A new EM algorithm for feature selection and model estimation is proposed.

• The algorithm is compared to other state-of-the-art approaches.

• Similar good or better results are obtained by the proposed method.

• Regarding time-critical applications and complex data, it outperforms other methods.

摘要

Highlights•A new EM algorithm for feature selection and model estimation is proposed.•The algorithm is compared to other state-of-the-art approaches.•Similar good or better results are obtained by the proposed method.•Regarding time-critical applications and complex data, it outperforms other methods.

论文关键词：Gaussian mixture models,Clustering,Feature selection,Feature saliency,Expectation maximization,Supervised learning,Remote sensing

论文评审过程：Received 27 November 2012, Revised 14 February 2014, Accepted 26 February 2014, Available online 6 March 2014.

论文官网地址：https://doi.org/10.1016/j.patcog.2014.02.015