Joint sparse regularization based Sparse Semi-Supervised Extreme Learning Machine (S3ELM) for classification

作者：

Highlights：

• A joint sparse regularizer is employed to prune ELM.

• A semi-supervised strategy is used to exploit the information of unlabeled samples.

• S3ELM is proposed to solve the pruning ELM model.

• The proof of the convergence of S3ELM is shown.

摘要

•A joint sparse regularizer is employed to prune ELM.•A semi-supervised strategy is used to exploit the information of unlabeled samples.•S3ELM is proposed to solve the pruning ELM model.•The proof of the convergence of S3ELM is shown.

论文关键词：Sparse semi-supervised learning,Extreme learning machine,ℓ2,1-Norm,Joint sparse regularization,Laplacian

论文评审过程：Received 19 January 2014, Revised 18 September 2014, Accepted 27 September 2014, Available online 13 October 2014.

论文官网地址：https://doi.org/10.1016/j.knosys.2014.09.014