“Noise Handling Capabilities of Multivariate Calibration Methods”

Authors: Rolf Ergon,
Affiliation: Telemark University College
Reference: 2002, Vol 23, No 4, pp. 259-273.

Keywords: PCR, PLSR, noise, prediction, spectra

Abstract: The noise handling capabilities of principal component regression (PCR) and partial least squares regression (PLSR) are somewhat disputed issues, especially regarding regressor noise. In an attempt to indicate an answer to the question, this article presents results from Monte Carlo simulations assuming a multivariate mixing problem with spectroscopic data. Comparisons with the best linear unbiased estimator (BLUE) based on Kalman filtering theory are included. The simulations indicate that both PCR and PLSR perform comparatively well even at a considerable regressor noise level. The results are also discussed in relation to estimation of pure spectra for the mixing constituents, i.e. to identification of the data generating system. In this respect solutions to well-posed least squares problems serve as references.

PDF PDF (1587 Kb)        DOI: 10.4173/mic.2002.4.2

DOI forward links to this article:
[1] Rolf Ergon, Maths Halstensen and Kim H. Esbensen (2011), doi:10.1002/cem.1356
[2] Maryam Ghadrdan, Chriss Grimholt and Sigurd Skogestad (2013), doi:10.1021/ie400542n
[3] Bernt Lie, David Di Ruscio, Rolf Ergon, Bjørn Glemmestad, Maths Halstensen, Finn Haugen, Saba Mylvaganam, Nils-Olav Skeie and Dietmar Winkler (2009), doi:10.4173/mic.2009.3.4
[4] Rolf Ergon (2013), doi:10.1002/9781118434635.ch08
[5] Rolf Manne, Randy J. Pell and L. Scott Ramos (2009), doi:10.1002/cem.1181
[6] R. Ergon (2009), doi:10.1002/cem.1180
[7] Rolf Ergon (2013), doi:10.1002/9781118434635.ch8
[1] BERNTSEN, H. (1988). Utvidet Kalmanfilter og multivariabel kalibrering, Report STF48 A88019, SINTEF, Trondheim, Norway.
[2] DI RUSCIO, D. (2000). A weighted view on the partial least squares algorithm, Automatica, 36, pp. 831-850 doi:10.1016/S0005-1098(99)00210-1
[3] ERGON, R. (1998). Dynamic system multivariate calibration by system identification methods, Modeling, Identification and Control, 19, No. 2, pp. 77-97 doi:10.4173/mic.1998.2.2
[4] ERGON, R. ESBENSEN, K.H. (2001). A didactically motivated MS prediction algorithm, Modeling. Identification and Control, 22, No. 3, pp. 131-139.
[5] FEARN, T. (2000). On orthogonal signal correction, Chemometrics Intell. Lab. Syst., 44, pp. 229-244.
[6] GELB, A. (1974). Applied Optimal Estimation, MIT Press, Mass.
[7] GREWAL, M.S. ANDREWS, A.P. (1993). Kalman Filtering: Theory and Practice, Prentice Hall, New Jersey.
[8] HELLAND, I.S. (1988). On the structure of partial least squares regression, Communications in statistics, 17, pp. 581-607.
[9] JOHNSON, R.A. WICHERN, D.W. (1998). Applied Multivariate Statistical Analysis, Prentice Hall, New Jersey.
[10] MANNE, R. (1987). Analysis of two partial-least-squares algorithms for multivariate calibration, Chemometrics Intell. Lab. Syst., 2, pp. 187-197 doi:10.1016/0169-7439(87)80096-5
[11] MARTENS, H. NÆS, T. (1989). Multivariate Calibration, Wiley, New York.
[12] TRYGG, J. WOLD, S. (2001). Orthogonal projections to latent structures, O-PLS. J. Chemometrics, 15, pp. 1-18.
[13] TRYGG, J. (2001). Parsimonious Multivariate Models, Ph.D. thesis, Umeå University, Department of Chemistry, Sweden.
[14] WESTERHUS. J.A., DE JONG, S., SMILDE, A.K. (2001). Direct orthogonal signal correction, Chemometrics Intell. Lab. Syst., 56, pp. 13-25 doi:10.1016/S0169-7439(01)00102-2
[15] WOLD, S., MARTENS, H. WOLD, H. (1982). The multivariate calibration problem in chemistry solved by the PLS method, Proc. Conf. Matrix pencils.
[16] WOLD, S., ANTTI, H., LINDGREN, F. OHMAN, J. (1988). Orthogonal signal correction of near-infrared spectra, Chemometrics Intell. Lab. Syst., 44, pp. 175-185 doi:10.1016/S0169-7439(98)00109-9

  title={{Noise Handling Capabilities of Multivariate Calibration Methods}},
  author={Ergon, Rolf},
  journal={Modeling, Identification and Control},
  publisher={Norwegian Society of Automatic Control}