基于ISOMAP降维的复杂轮廓异常点识别方法

摘要/Abstract

摘要： 高维复杂轮廓异常点识别方法研究是目前过程轮廓监控的重要课题之一。以高维复杂轮廓为研究对象，建立非参数轮廓矩阵模型，将基于测地距离的ISOMAP非线性降维技术与χ2控制图相结合，提出新的轮廓异常点识别方法，以实现高维复杂轮廓异常点的准确识别。仿真实验和实际案例的应用分析结果证实该方法在异常点识别的准确性方面具有良好的性能。

关键词: 异常点识别, 等距特征映射（ISOMAP）, 轮廓, 降维

Abstract: Researches of outlierdetection methods for the highdimensional complex profiles are as an important branch of profile monitoring. Considering the complexity of some highdimensional profiles, a matrix model for nonparametric profile was set up. Based on this model, a combined method was proposed, where ISOMAP could reduce the dimensionality based on geodesic distance and χ2 control chart could identify profile outliers. The proposed method shows good performance in the accuracy of outlier detection by both simulation studies and real case studies.

Key words: outlier detection, isometric mapping(ISOMAP), profile, dimensiondeduction

中图分类号:

TH165.4

聂斌, 李京亚, 姚雪海. 基于ISOMAP降维的复杂轮廓异常点识别方法[J]. 中国机械工程.

Nie Bin, Li Jingya, Yao Xuehai. Detecting Outliers in Complex Profiles with ISOMAP[J]. China Mechanical Engineering.

参考文献

[1]Woodall W H, Spitzner D J, Montgomery D C, et al. Using Control Charts to Monitor Process and Product Quality Profiles[J]. Journal of Quality Technology, 2004, 36(3): 309-320.

[2]Hawkins D M, Peihua Q, Chang W K. The Changepoint Model for Statistical Process Control[J]. Journal of Quality Technology, 2003, 35(4): 355-366.

[3]Koh C K H, Shi J, Black J M, et al. Tonnage Signature Attribute Analysis for Stamping Process[J]. Transactions—North American Manufacturing Research Institution of SME, 1996, 24: 193-198.

[4]Walker E, Wright S P. Comparing Curves Using Additive Models[J]. Journal of Quality Technology, 2002, 34(1): 118-129.

[5]Tracy N D. Multivariate Control Charts for Individual Observations[J]. Journal of Quality Technology, 1992, 24(2): 88-95.

[6]Kang L, Albin S. On-line Monitoring When the Process Yields a Linear Profile[J]. Journal of Quality Technology, 2000, 32(4): 418-426.

[7]Nagappan N, Williams L, Vouk M, et al. Using In-process Testing Metrics to Estimate Post-release Field Quality[C]// ISSRE07. The 18th IEEE International Symposium on Software Reliability, 2007. Trollhattan: IEEE, 2007: 209-214.

[8]Zhang H, Albin S. Detecting Outliers in Complex Profiles Using a χ2 Control Chart Method[J]. IIE Transactions, 2009, 41(4): 335-345.

[9]Chang S I, Yadama S. Statistical Process Control for Monitoring Non-linear Profiles Using Wavelet Filtering and B-spline Approximation[J]. International Journal of Production Research, 2010, 48(4): 1049-1068.

[10]Girimurugan S, Chicken E, Pignatiello Jr J J, et al. Wavelet Anova for Detection of Local and Global Profile Changes[C]//Proceedings of the 2013 Industrial and Systems Engineering Research Conference. San Juan, Puerto Rico:IIE. 2013: 3235-3244.

[11]Ding Y, Zeng L, Zhou S. Phase I Analysis for Monitoring Nonlinear Profiles in Manufacturing Processes[J]. Journal of Quality Technology, 2006, 38(3): 199-216.

[12]Colosimo B M, Pacella M. On the Use of Principal Component Analysis to Identify Systematic Patterns in Roundness Profiles[J]. Quality and Reliability Engineering International, 2007, 23(6): 707-725.

[13]Shiau J J H, Huang H L, Lin S H, et al. Monitoring Nonlinear Profiles with Random Effects by Nonparametric Regression[J]. Communications in Statistics—Theory and Methods, 2009, 38(10): 1664-1679.

[14]Lee J A, Verleysen M. Nonlinear Dimensionality Reduction[M]. New York: Springer Science & Business Media, 2007.

[15]Neto F M, de Magalhes M S. A Laplacian Spectral Method in Phase I Analysis of Profiles[J]. Applied Stochastic Models in Business and Industry, 2012, 28(3): 251-263.

[16]Tenenbaum J B, de Silva V, Langford J C. A Global Geometric Framework for Nonlinear Dimensionality Reduction[J]. Science, 2000, 290(5500): 2319-2323.

[17]Mardia K V, Kent J T, Bibby J M. Multivariate Analysis [M]. London: Academic Press, 1979.

[18]Samko O, Marshall A D, Rosin P L. Selection of the Optimal Parameter Value for the Isomap Algorithm[J]. Pattern Recognition Letters, 2006, 27(9): 968-979.