Support Vector Machines Approach for Detecting the Mean Shifts in Hotelling-s T2 Control Chart with Sensitizing Rules
In many industries, control charts is one of the most
frequently used tools for quality management. Hotelling-s T2 is used
widely in multivariate control chart. However, it has little defect when
detecting small or medium process shifts. The use of supplementary
sensitizing rules can improve the performance of detection. This study
applied sensitizing rules for Hotelling-s T2 control chart to improve the
performance of detection. Support vector machines (SVM) classifier
to identify the characteristic or group of characteristics that are
responsible for the signal and to classify the magnitude of the mean
shifts. The experimental results demonstrate that the support vector
machines (SVM) classifier can effectively identify the characteristic
or group of characteristics that caused the process mean shifts and the
magnitude of the shifts.
[1] F.Aparisi, C.W. Champ, and J.C. García-Díaz, "A Performance Analysis
of Hotelling's
¤ç 2 Control Chart with Supplementary Runs Rules,"
Quality Engineering, vol. 16, pp.359-368, 2004.
[2] J. E. Jackson, "Multivariate Quality Control Communications in Statistics:
Theory and Methods" vol. 14, pp. 2657-2688,1985.
[3] R.L.Mason, N.D. Tracy, and J.C. Young, "Decomposition of T2 for
multivariate control chart interpretation," Journal of Quality Technology,
vol.27, pp.99-108, 1995.
[4] G.A. Pugh, "A comparison of neural networks to SPC charts,"
International journal of Production Research, vol.21, pp.253-255, 1991.
[5] T.-Y.Wang, and L.-H.Chen, "Mean shifts detection and classification in
multivariate process: a neural-fuzzy approach," Journal of Intelligent
Manufacturing, vol. 13, pp.211-221, 2002.
[6] L.- H.Chen, and T.- Y. Wang, " Artificial neural networks to classify
mean shifts from multivariate
¤ç 2 chart signals," Computers & Industrial
Engineering, vol.47, pp.195-205, 2004.
[7] R.B.Chinnam, "Support vector machines for recognizing shifts in
correlated and other manufacturing processes," International Journal of
Production Research, vol. 40, pp.4449-4466, 2002.
[1] F.Aparisi, C.W. Champ, and J.C. García-Díaz, "A Performance Analysis
of Hotelling's
¤ç 2 Control Chart with Supplementary Runs Rules,"
Quality Engineering, vol. 16, pp.359-368, 2004.
[2] J. E. Jackson, "Multivariate Quality Control Communications in Statistics:
Theory and Methods" vol. 14, pp. 2657-2688,1985.
[3] R.L.Mason, N.D. Tracy, and J.C. Young, "Decomposition of T2 for
multivariate control chart interpretation," Journal of Quality Technology,
vol.27, pp.99-108, 1995.
[4] G.A. Pugh, "A comparison of neural networks to SPC charts,"
International journal of Production Research, vol.21, pp.253-255, 1991.
[5] T.-Y.Wang, and L.-H.Chen, "Mean shifts detection and classification in
multivariate process: a neural-fuzzy approach," Journal of Intelligent
Manufacturing, vol. 13, pp.211-221, 2002.
[6] L.- H.Chen, and T.- Y. Wang, " Artificial neural networks to classify
mean shifts from multivariate
¤ç 2 chart signals," Computers & Industrial
Engineering, vol.47, pp.195-205, 2004.
[7] R.B.Chinnam, "Support vector machines for recognizing shifts in
correlated and other manufacturing processes," International Journal of
Production Research, vol. 40, pp.4449-4466, 2002.
@article{"International Journal of Mechanical, Industrial and Aerospace Sciences:54168", author = "Tai-Yue Wang and Hui-Min Chiang and Su-Ni Hsieh and Yu-Min Chiang", title = "Support Vector Machines Approach for Detecting the Mean Shifts in Hotelling-s T2 Control Chart with Sensitizing Rules", abstract = "In many industries, control charts is one of the most
frequently used tools for quality management. Hotelling-s T2 is used
widely in multivariate control chart. However, it has little defect when
detecting small or medium process shifts. The use of supplementary
sensitizing rules can improve the performance of detection. This study
applied sensitizing rules for Hotelling-s T2 control chart to improve the
performance of detection. Support vector machines (SVM) classifier
to identify the characteristic or group of characteristics that are
responsible for the signal and to classify the magnitude of the mean
shifts. The experimental results demonstrate that the support vector
machines (SVM) classifier can effectively identify the characteristic
or group of characteristics that caused the process mean shifts and the
magnitude of the shifts.", keywords = "Hotelling's T2 control chart, Neural networks, Sensitizing rules, Support vector machines.", volume = "7", number = "6", pages = "1101-5", }