摘要
针对传统极端学习机算法(ELM)和K近邻分类算法(KNN)在处理分类问题中存在的问题,提出一种基于PSOELM特征映射的KNN分类算法。该算法利用ELM的输入层权值和隐层神经元对输入样本进行非线性映射,并利用粒子群算法(PSO)寻找一组最优的ELM映射参数,再将映射后的特征样本输入到KNN算法中,提高处理线性不可分问题的能力。在多个数据集上的实验结果表明,文中算法比KNN改进算法以及ELM改进算法有更高的分类正确率。
The K-nearest neighbor(KNN)classification algorithm based on PSO.ELM(particle swarm optimization-extreme learning machine)feature mapping is proposed because the traditional ELM algorithm and KNN classification algorithm have some shortcomings in classification process.The input-layer weight and hidden.layer neuron of ELM algorithm are used to perform the nonlinear mapping for the input sample.The PSO algorithm is used to find a group of optimal ELM mapping parameters,and then the mapped feature sample is input into KNN algorithm,which can improve the ability to deal with the linear impartibility problem.The experimental results of several data sets show that the proposed algorithm has higher classification accuracy than improved KNN algorithm and improved ELM algorithm.
作者
丁建立
刘涛
王家亮
曹卫东
DING Jianli;LIU Tao;WANG Jialiang;CAO Weidong(College of Computer Science and Technology,Civil Aviation University of China,Tianjin 300300,China;Tianjin Key Lab for Advanced Signal Processing,Civil Aviation University of China,Tianjin 300300,China)
出处
《现代电子技术》
北大核心
2019年第5期152-156,共5页
Modern Electronics Technique
基金
民航局科技创新引导资金专项(MHRD20150107)
中国民航大学天津市智能信号与图像处理重点实验室开放基金(2015ASP02)~~
关键词
K近邻分类算法
极端学习机
特征映射
粒子群算法
混合算法
线性不可分
K-nearest neighbor classification algorithm
extreme learning machine
feature mapping
particle swarm optimization algorithm
hybrid algorithm
linear impartibility
