A comparison of multivariate calibration techniques applied to experimental NIR data sets - Part III: Robustness against instrumental perturbation conditions

被引:11
作者
Estienne, F
Despagne, F
Walczak, B
de Noord, OE
Massart, DL
机构
[1] Free Univ Brussels, ChemoAC, Farmaceut Inst, B-1090 Brussels, Belgium
[2] Shell Int Chem BV, Shell Res & Technol Ctr Amsterdam, NL-1030 BN Amsterdam, Netherlands
关键词
multivariate calibration; method comparison; instrumental change; extrapolation; nonlinearity; clustering;
D O I
10.1016/j.chemolab.2004.04.007
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
This work is part of a more general research aiming at comparing the performance of multivariate calibration methods. In the first and second parts of the study, the performances of multivariate calibration methods were compared in situations of interpolation and extrapolation, respectively. This third part of the study deals with robustness of calibration methods in the case where spectra corresponding to new samples of which the y value has to be predicted are affected by instrumental perturbations not accounted for in the calibration set. These types of perturbations can happen due to instrument ageing, replacement of one or several parts of the spectrometer (e.g. the detector), use of a new instrument, or modifications in the measurement conditions, like the displacement of the instrument to a different location. Although no general rules could be extracted from the results, the variety of data sets and methods tested allowed some guidelines for multivariate calibration in this unfavourable case of instrumental perturbation to be given. Models based on Neural Networks (NN) applied to Fourier Coefficient proved particularly robust in some cases, and failed badly in some others. The most stable methods were principal component regression (PCR, with component selection) and partial least squared regression (with complexity optimisation performed by randomisation test). (C) 2004 Elsevier B.V. All rights reserved.
引用
收藏
页码:207 / 218
页数:12
相关论文
共 23 条
[1]  
[Anonymous], 1989, GENETIC ALGORITHM SE
[2]  
[Anonymous], 1989, MULTIVARIATE CALIBRA
[3]   Elimination of uninformative variables for multivariate calibration [J].
Centner, V ;
Massart, DL ;
deNoord, OE ;
deJong, S ;
Vandeginste, BM ;
Sterna, C .
ANALYTICAL CHEMISTRY, 1996, 68 (21) :3851-3858
[4]   Optimization in locally weighted regression [J].
Centner, V ;
Massart, DL .
ANALYTICAL CHEMISTRY, 1998, 70 (19) :4206-4211
[5]   Comparison of multivariate calibration techniques applied to experimental NIR data sets [J].
Centner, V ;
Verdú-Andrés, J ;
Walczak, B ;
Jouan-Rimbaud, D ;
Despagne, F ;
Pasti, L ;
Poppi, R ;
Massart, DL ;
de Noord, OE .
APPLIED SPECTROSCOPY, 2000, 54 (04) :608-623
[6]   SIMPLS - AN ALTERNATIVE APPROACH TO PARTIAL LEAST-SQUARES REGRESSION [J].
DEJONG, S .
CHEMOMETRICS AND INTELLIGENT LABORATORY SYSTEMS, 1993, 18 (03) :251-263
[7]   Variable selection for neural networks in multivariate calibration [J].
Despagne, F ;
Massart, DL .
CHEMOMETRICS AND INTELLIGENT LABORATORY SYSTEMS, 1998, 40 (02) :145-163
[8]   Optimization of partial-least-squares calibration models by simulation of instrumental perturbations [J].
Despagne, F ;
Massart, DL ;
deNoord, OE .
ANALYTICAL CHEMISTRY, 1997, 69 (16) :3391-3399
[9]  
Draper N. R., 1966, APPL REGRESSION ANAL
[10]   A comparison of multivariate calibration techniques applied to experimental NIR data sets Part II. Predictive ability under extrapolation conditions [J].
Estienne, F ;
Pasti, L ;
Centner, V ;
Walczak, B ;
Despagne, F ;
Rimbaud, DJ ;
de Noord, OE ;
Massart, DL .
CHEMOMETRICS AND INTELLIGENT LABORATORY SYSTEMS, 2001, 58 (02) :195-211