曲阜师范大学学报（自然科学版）2024，Vol.50Issue(4)：53-58,6.DOI:10.3969/j.issn.1001-5337.2024.4.053
非最小均方误差下的核主成分分析算法
Kernel principal component analysis algorithm under non-minimum mean square error
摘要
Abstract
The traditional machine learning method has a better processing effect on low-dimensional data,but it is often not satisfactory for high-dimensional nonlinear data.For high-dimensional nonlinear da-ta processing,this paper improves a novel machine learning algorithm which can eliminate the redundant and irrelevant features in high-dimensional data and reduce the dimension of these nonlinear high-dimen-sional data,namely kernel principal component analysis algorithm based on maximum correntropy criterion(KPCA-MCC).In this paper,the particle swarm optimization algorithm is used to optimize the data param-eters.Under the support of vector machine(SVM)and least squares vector machine(LSSVM),the wind power is simulated and predicted.The KPCA-SVM and PCA-SVM,KPCA-MCC and KPCA-MSE algo-rithms are compared.Through a large number of comparative experiments,it is proved that the proposed kernel principal component analysis algorithm based on maximum correntropy criterion is effective,appli-cable and robust.关键词
主成分分析/核主成分分析/最大相关熵/鲁棒性/风电功率Key words
principal component analysis/kernel principal component analysis/maximum correlation entropy/robustness/wind power分类
信息技术与安全科学引用本文复制引用
李建磊,付世豪,刘志鹏..非最小均方误差下的核主成分分析算法[J].曲阜师范大学学报（自然科学版）,2024,50(4):53-58,6.基金项目
河南省自然科学基金(222300420579) （222300420579）
河南省高等学校青年骨干教师培养计划(2019GGJS100) （2019GGJS100）
河南省高等学校重点科研项目计划基础研究专项(20zx003) （20zx003）
河南省研究生优质课程(YJS2022KC01). （YJS2022KC01）
