Title :
Kernel Principal Component Chart for Defect Detection
Author :
Stefatos, George ; Luo, Yan ; Hamza, A. Ben
Author_Institution :
Concordia Univ., Montreal
Abstract :
We present a multivariate statistical process control chart using kernel principal component analysis. The proposed control chart is robust to outliers, and its control limits are derived from the eigenanalysis of the Gaussian kernel matrix in the Hilbert feature space. Our experimental results show that the much improved performance of the proposed control in comparison with existing monitoring and controlling charts.
Keywords :
Gaussian processes; Hilbert spaces; control charts; eigenvalues and eigenfunctions; matrix algebra; multivariable systems; principal component analysis; robust control; statistical process control; Gaussian kernel matrix; Hilbert feature space; controlling chart; defect detection; eigenanalysis; kernel principal component chart; monitoring chart; multivariate statistical process control chart; robust control; Control charts; Covariance matrix; Hilbert space; Kernel; Monitoring; Principal component analysis; Process control; Quality control; Robust control; Statistics;
Conference_Titel :
Electrical and Computer Engineering, 2007. CCECE 2007. Canadian Conference on
Conference_Location :
Vancouver, BC
Print_ISBN :
1-4244-1020-7
Electronic_ISBN :
0840-7789
DOI :
10.1109/CCECE.2007.154
