Scholarlink

<p>In view of the simple structure of transfer function, an integrated strategy of transfer function logic operation is proposed, which combines different transfer functions reasonably and improves the classification accuracy. It is applied to the binary Pelican optimization algorithm (BPOA) to solve the high-dimensional feature selection problem. At the same time, the adaptive retractable strategy and the electric eel hunting strategy are introduced to improve the Pelican optimization algorithm (POA) in order to improve the unbalanced exploration and exploitation ability of the Pelican optimization algorithm itself, which is easy to fall into local optimal problems. The simulation experiment is divided into three parts. Firstly, two different logical operation integration routes are proposed, and four different transfer functions of S-shaped, V-shaped, U-shaped and RZ-shaped are respectively integrated, and the POA variant with the best effect is selected. In the following experiment, the transfer function of the same integration idea is combined in pairs to further improve the integrated transfer function of logical operations. Finally, adaptive scaling strategy and electric eel hunting strategy are introduced to improve the existing problems of POA, and compared with BAOA, BCOA, BEO, BHHO, BPDO and BZOA. Twelve standard high-dimensional UCI datasets were used to perform performance tests, and the results were statistically analyzed by Friedman test and Wilcoxon rank sum test. Simulation results show that the improved method can effectively simplify feature subsets, improve classification accuracy and obtain lower fitness values in solving these data sets.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Multi-strategy fusion pelican optimization algorithm and logic operation ensemble of transfer functions for high-dimensional feature selection problems

  • 期刊论文

  • 发表日期：

    2025年1月17日

查看全文
  • Hao-Ming Song,
  • Jie-Sheng Wang,
  • Jia-Ning Hou,
  • Yu-Cai Wang,
  • Yu-Wei Song,
  • Yu-Liang Qi

摘要

In view of the simple structure of transfer function, an integrated strategy of transfer function logic operation is proposed, which combines different transfer functions reasonably and improves the classification accuracy. It is applied to the binary Pelican optimization algorithm (BPOA) to solve the high-dimensional feature selection problem. At the same time, the adaptive retractable strategy and the electric eel hunting strategy are introduced to improve the Pelican optimization algorithm (POA) in order to improve the unbalanced exploration and exploitation ability of the Pelican optimization algorithm itself, which is easy to fall into local optimal problems. The simulation experiment is divided into three parts. Firstly, two different logical operation integration routes are proposed, and four different transfer functions of S-shaped, V-shaped, U-shaped and RZ-shaped are respectively integrated, and the POA variant with the best effect is selected. In the following experiment, the transfer function of the same integration idea is combined in pairs to further improve the integrated transfer function of logical operations. Finally, adaptive scaling strategy and electric eel hunting strategy are introduced to improve the existing problems of POA, and compared with BAOA, BCOA, BEO, BHHO, BPDO and BZOA. Twelve standard high-dimensional UCI datasets were used to perform performance tests, and the results were statistically analyzed by Friedman test and Wilcoxon rank sum test. Simulation results show that the improved method can effectively simplify feature subsets, improve classification accuracy and obtain lower fitness values in solving these data sets.