A Robust Probability Principle Component Analysis Method

穆向阳 1; 2; 张太镒 1; 周亚同 3

您当前的位置：

首页 >

文章列表页 >

A Robust Probability Principle Component Analysis Method

更新时间：2025-12-08

- A Robust Probability Principle Component Analysis Method
- Vol. 42, Issue 10, Pages: 1217-1220(2008)
- 作者机构：
  
  1. 西安交通大学电子与信息工程学院,西安,710049
  2. 西安石油大学电子工程学院,西安,710065
- 作者简介：
- 基金信息：
- DOI：
  CLC： TP18
- Online First：10 October 2008，
  
  Published：2008
- 稿件说明：
移动端阅览
穆向阳 1, 2, 张太镒 1, et al. A Robust Probability Principle Component Analysis Method[J]. 2008, 42(10): 1217-1220.
DOI：

穆向阳 1, 2, 张太镒 1, et al. A Robust Probability Principle Component Analysis Method[J]. 2008, 42(10): 1217-1220. DOI：

摘要

针对传统主成分对实际样本的奇点不敏感的缺陷

提出了一种鲁棒概率主成分分析(RPPCA)方法.首先引入连续的决策变量构造新能量函数

将事先给定的硬门限改为自适应确定的软门限

门限值由样本自动确定

再计算概率主成分进行特征提取.与线性主成分分析(LPCA)和概率主成分分析(PPCA)方法相比

RPPCA方法更为实用

有效地减小了奇点的影响

显示出比PPCA更强的稳健性

也扩大了实用范围.实验结果表明

RPPCA方法的分类准确率比LPCA方法平均提高了3.2%

比PPCA方法平均提高了0.7%.

Abstract

In order to overcome the drawback that traditional principal component analysis fails to the outliers existing in the realistic data

a robust probability principal component analysis(RPPCA)method is proposed. A continuous decision variable is introduced into the energy function

and the preset hard threshold is replaced by a soft adaptive threshold which is automatically determined by the data. The algorithm is then embedded in the procedure of PPCA's principal component feature extraction. Compared with PCA and probability principal component analysis(PPCA)

the proposed RPPCA can resist outlier well

is more robust than PPCA

and enlarges the real application area.The simulation results show that the algorithm improves 3.2% classification accuracy to LPCA

and 0.7% to PPCA on average.

关键词

Keywords

references

TIPPING M E,BISHOP C M. Probabilistic principal component analysis [J]. J Roy Statist Soc, Series B,1999, 61(3): 611-622.

BISHOP C M. Latent variable models, learning in graphical models [M]. Jordan M I. Cambridge, MA, USA: MIT Press, 1999:371-403.

ZHAO Weixiang, CHEN Dezhao, HU Shangxu. Detection of outlier and a robust BP algorithm against outlier [J]. Computers Chemical Engineering, 2004, 28(8): 1403-1408.

BULLEN R J, CORNFORD D, NABNEY I T. Outlier detection in scatterometer data: neural network approaches [J]. Neural Networks, 2003, 16(3): 419-426.

XU Lei, YUILLE A L. Robust principal component analysis by self-organizing rules based on statistical physics approach [J]. IEEE Trans on Neural Networks, 1995, 6(1): 131-143.

GAVRILA D M, GIEBEL J. Shape-based pedestrian detection and tracking [C]∥ Proceedings of IEEE Intelligent Vehicle Symposium. Piscataway, NJ, USA:IEEE Press, 2002: 8-14.

Zhao L, Thorpe C E. Stereo and neural network-based pedestrian detection [J]. IEEE Trans on Intelligent Transportation Systems, 2000, 1(3): 148-154.

The MIT-CBCL Face Recognition Database [EB/OL].(2003-01-08)[2007-11-16].http:∥cbcl.mit.edu/software-datasets/heisele/facerecognition-database.html.

CRISTIANINI N, SHAWE-TAYLOR J. An introduction to support vector machines and other kernel-based learning methods [M]. Cambridge, England: Cambridge University Press, 2000.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Research and Application of Ensemble Empirical Mode Decomposition Principle and 1.5 Dimension Spectrum Method

Nonlinear Manifold Learning Method of Mechanical Impact Faults Extraction

Feature Extraction for Incipient Fault Diagnosis of Rolling Bearings Based on Nonlinear Manifold Learning

A Novel Approach of Feature Extraction for Advanced Radar Emitter Signals Using Time-Frequency Atom Decomposition

A Fast Implementation of Null Space Based Algorithm

Related Author

成玮 1

何正嘉 1

訾艳阳 1

陈略 1

梁小影 2

张熠卓 2

栗茂林 2

Related Institution

Beijing Aerospace Control Center

State Key Laboratory for Manufacturing Systems Engineering, Xi'an Jiaotong University

School of Mechanical Engineering, Xi'an Jiaotong University

School of Information Science and Technology, Southwest Jiaotong University

School of Computer Science and Technology, Nanjing University of Science and Technology

⁰