Sequential random k-nearest neighbor feature selection for high-dimensional data

作者：

Highlights：

• The study proposes the ensemble-based feature selection algorithm.

• The proposed algorithm is especially useful for large p and small n problems.

• Experiments with 20 real data demonstrated the effectiveness of the proposed algorithm.

摘要

•The study proposes the ensemble-based feature selection algorithm.•The proposed algorithm is especially useful for large p and small n problems.•Experiments with 20 real data demonstrated the effectiveness of the proposed algorithm.

论文关键词：Feature selection,High dimensionality,Ensemble,Wrapper,Random forest,k-NN

论文评审过程：Available online 1 November 2014.

论文官网地址：https://doi.org/10.1016/j.eswa.2014.10.044