An adaptive length-variation based evolutionary multitasking algorithm for feature selection of high-dimensional classification

Lingjie Li, Yuze Zhang, Zhijiao Xiao, Qiuzhen Lin, Xin Wang, Xiuqiang He, Zhong Ming

Abstract: Highlights•An evolutionary multitasking approach with adaptive length variation is designed in this paper.•A multitasking construction strategy based on relevance and adaptive threshold is designed to construct two subtasks.•A competitive swarm optimizer is used to facilitate efficient knowledge transfer between the subtasks.•A variable-length individual initialization scheme based on Gaussian distribution is designed to generate high-quality initial particle lengths.•An adaptive length variation scheme is designed to dynamically adjust the particle lengths.