计算机应用 ›› 2013, Vol. 33 ›› Issue (04): 998-1000.DOI: 10.3724/SP.J.1087.2013.00998

• 人工智能 • 上一篇    下一篇

基于最大间隔超平面的增强特征提取算法

侯勇1,2,郑雪峰1   

  1. 1. 北京科技大学 计算机与通信工程学院,北京 100083
    2. 山东经贸职业学院 科学与人文学院,山东 潍坊 261011
  • 收稿日期:2012-10-19 修回日期:2012-11-30 出版日期:2013-04-01 发布日期:2013-04-23
  • 通讯作者: 侯勇
  • 作者简介:侯勇(1978-),男,山东蓬莱人,讲师,博士研究生,主要研究方向:数据挖掘、网络安全、机器学习;郑雪峰(1951-),男,福建福州人,教授,主要研究方向:网络安全。

Margin maximizing hyperplanes based enhanced feature extraction algorithm

HOU Yong1,2,ZHENG Xuefeng2   

  1. 1. College of Humanities and Science, Shandong Vocational College of Economics and Business, Weifang Shandong 261011, China
    2. School of Computer and Communication Engineering, University of Science and Technology Beijing, Beijing 100083 China
  • Received:2012-10-19 Revised:2012-11-30 Online:2013-04-01 Published:2013-04-23
  • Contact: HOU Yong

摘要: 核主成分分析(KPCA)与多层感知器(MLP)是流行的特征提取算法,但这些算法存在效率低下与易陷于局部最优解等问题。针对KPCA与MLP算法存在的问题,提出了一个新颖的特征提取算法——基于最大间隔超平面的增强的特征提取算法(EFE)。该算法独立于输入样本的概率分布,通过采用隔间最大化且两两正交的最大分割超平面,将输入样本映射到超平面的法线所张成的子空间中,实现输入样本的特征提取。在对现实世界数据集wine与AR的特征提取的实验表明,基于最大间隔超平面的增强特征提取算法在执行效率、识别准确率方面均超出了KPCA与MLP的执行效率与识别准确率。

关键词: 特征提取, 降维, 核主成分分析, 多层感知器, 最大间隔超平面, 内在维数

Abstract: Kernel Principal Component Analysis (KPCA) and Multi-Layer Perceptron (MLP) neural network are popular feature extraction algorithms. However, these algorithms are inefficient and easy to fall into local optimal solution. The paper proposed a new feature extraction algorithm — margin maximizing hyperplanes based Enhanced Feature Extraction algorithm (EFE), which can overcome the problem of KPCA and MLP algorithm. The proposed EFE algorithm, whcih maps the input samples to the subspace spanned by the normals of hyperplanes through adopting the pairwise orthogonal margin maximizing hyperplanes, is independent of the probability distribution of the input samples. The results of these feature extraction experiments on real world data set — wine and AR show that FE algorithm is beyond KPCA and MLP in terms of the efficiency of the implementation and accuracy of recognition.

Key words: feature extraction, dimensionality reduction, Kernel Principal Component Analysis (KPCA), Multi-Layer Perceptron (MLP), Margin maximizing hyperplanes, intrinsic dimension

中图分类号: