ORF  >> Vol. 1 No. 1 (August 2011)

    基于SOM特征映射空间相似度判别的软传感器建模变量选择
    Variable Selection for Soft-Sensing Model Based on False Nearest Neighbors in Self-Organizing Feature Mapping Feature Space

  • 全文下载: PDF(354KB) HTML    PP.16-21   DOI: 10.12677/orf.2011.11004  
  • 下载量: 2,768  浏览量: 9,733   国家自然科学基金支持

作者:  

侯杰,李太福,余德君,程杨

关键词:
变量选择软传感器建模SOM神经网络特征空间虚假最近邻点法
ariable Selection; Soft-Sensoring Modeling; Self-Organizing Feature Mapping (SOM); Feature Space; False Nearest Neighbor (FNN)

摘要:

针对软传感器建模中存在的信息冗余,提出一种基于自组织特征映射神经网络(Self-Organizing Feature Mapping,SOM)的变量选择方法。该方法借助SOM简单快速的特征映射能力对数据进行投影,采用虚假最近邻点法(False Nearest Neighbor,FNN)计算某变量删减前后数据在SOM投影空间的相似度,通过相似度来判断其对主导变量的解释能力,由此进行变量的选择。实验结果表明该方法能有效的进行变量选择,为软传感器建模变量选择提供了一种新思路。

A new variable selection method based on Self-Organizing Feature Mapping (SOM) is proposed for soft-sensing modeling to eliminate redundant information. In the proposed method, SOM is used to get new space from original variable space because of its simple and fast. The False Nearest Neighbor (FNN) is used to calculate the similarity of data in the new SOM space. The primary variable would be estimated to select secondary variables. The results show that the method is effective and suitable for variable selection. Therefore, a new method is provided for the variable selection of soft-sensing modeling.

文章引用:
侯杰, 李太福, 余德君, 程杨. 基于SOM特征映射空间相似度判别的软传感器建模变量选择[J]. 运筹与模糊学, 2011, 1(1): 16-21. http://dx.doi.org/10.12677/orf.2011.11004

参考文献

[1] 王孝红, 刘文光, 于宏亮. 工业过程软测量研究[J]. 济南大学学报: 自然科学版, 2009, 23(1): 80-86.
[2] M. Gevrey, I. Dimopoulos, and S. Lek. Review and comparison of methods to study the contribution of variables in artificial neural network models. Ecological Modelling, 2003, 160(3): 249 -264.
[3] Q. R. Chen, C. H. Yang. Quasi-stepwise regression variable selection and its application in rural household net income fore-casting. Systems Engineering: Theory & Practice, 2008, 28(11): 16-22.
[4] 贾洪飞, 隽志才, 王晓原等. 利用因子分析选取车辆跟驰模型输入变量[J]. 公路交通科技, 2004, 21(1): 81-84.
[5] K. O. Elish, M. O. Elish. Predicting defect-prone software mod-ules using support vector machines. The Journal of Systems and Software, 2008, 81(5): 649-660.
[6] F. Westad, M. Hersleth, P. Lea, et al. Variable selection in PCA in sensory descriptive and consumer data. Food Quality and Preference, 2003, 14(5-6): 463-472.
[7] W.-S. Lee, Y.-S. Kwon, J.-C. Yoo, et al. Multivariate analysis and self-organizing mapping applied to analysis of nest-site selec-tionin Black-tailed Gulls. Ecological Modelling, 2006, 193(3-4): 602-614.
[8] G. R. Lloyd, K. Wongravee, C. J. L. Silwood, et al. Self organis-ing maps for variable selection: Application to human saliva analyzed by nuclear magnetic resonance spectroscopy to inves-tigate the effect of an oral health care product. Chemometrics and Intelligent Laboratory Systems, 2009, 98(2): 149-161.
[9] F. Corona, E. Liitiainen, A. Lendasse, et al. A SOM-based ap-proach to estimating product properties from spectroscopic meas-urements. Neurocomputing, 2009, 73(1-3): 71-79.
[10] K. M. Najman, K. Najman. Applying the Kohonen self-orga-nizing map networks to select variables. Data Analysis, Machine Learning and Applications, 2008, 1: 45-54.
[11] Y.-S. Park, J. Tison, S. Lek, et al. Application of a self- organiz-ing map to select representative species in multivariate analysis: A case study determining diatom distribution patterns across France. Ecological Informatics, 2006, 1(3): 247-257.
[12] 廖广兰, 陈勇辉, 史铁林. 自组织映射网络的可视化研究[J].计算机工程与应用, 2003, 39(9): 35-37.
[13] 王海燕, 盛昭瀚. 混沌时间序列相空间重构参数的选取方法[J]. 东南大学学报(自科版), 2000, 30(5): 113-117.
[14] R. Philips, I. Guttman. A new criterion for variable selection. Statistics & Probability Letters, 1998, 38(1): 11-19.
[15] A. Eleuteri, R. Tagliaferri, and L. Milano. A novel information geometric approach to variable selection in MLP networks. Neural Networks, 2005, 18(10): 1309-1318.