Particle distance rank feature selection by particle swarm optimization
作者：
Highlights：
• A multi-objective particle swarm optimization feature selection is proposed.
• A new particle ranking is proposed based on dominated and non-dominated particles.
• A new feature ranking is proposed to update position and velocity of particles.
• Properties of the proposed method are proven mathematically.
• The proposed method outperforms state-of-the-art feature selection methods.
摘要
•A multi-objective particle swarm optimization feature selection is proposed.•A new particle ranking is proposed based on dominated and non-dominated particles.•A new feature ranking is proposed to update position and velocity of particles.•Properties of the proposed method are proven mathematically.•The proposed method outperforms state-of-the-art feature selection methods.
论文关键词：Feature selection,Particle ranking,Particle swarm optimization,Multi-objective optimization
论文评审过程：Received 18 August 2020, Revised 4 July 2021, Accepted 13 July 2021, Available online 26 July 2021, Version of Record 28 July 2021.
论文官网地址：https://doi.org/10.1016/j.eswa.2021.115620
