摘要:

AbstractBy means of Monte Carlo simulations a comparison has been made between ordinary least squares regression and robust regression. The robust regression procedure is based on the Huber estimate and is computed by means of the iteratively reweighted least squares algorithm. The performance of both procedures has been evaluated for estimation of the parameters of a calibration function and for determination of the concentration of unknown samples. The influence of the distributional characteristics skewness and kurtosis has been studied, and the number of measurements used for constructing the calibration curve has also been taken into account. Under certain conditions robust regression offers an advantage over least squares regression.

ISSN:0886-9383    

DOI:10.1002/cem.1180030203

出版商:John Wiley&Sons, Ltd.    

年代:1989

数据来源: WILEY

摘要:

AbstractA validation protocol for multicomponent spectroscopic assays based on principal components regression is described. Factorial design and hypothesis tests are used to establish the linearity and absence of interaction between components in the regression model. Testing considers multiple response variables simultaneously so that correlation between residuals is properly treated. Assay reproducibility and sensitivity to related substances are evaluated.

ISSN:0886-9383    

DOI:10.1002/cem.1180030204

出版商:John Wiley&Sons, Ltd.    

年代:1989

数据来源: WILEY

摘要:

AbstractThe theory of experimental error in analysis of mixture experiments by abstract factor analysis or target transformation factor analysis is considered. The theoretical implications of using signal‐to‐noise ratios (as weights) or canonical variates analysis to reduce the level of imbedded error in the factor model are examined. The approach is illustrated by application to13C NMR spectra of lubricant basestock mixtu

ISSN:0886-9383    

DOI:10.1002/cem.1180030205

出版商:John Wiley&Sons, Ltd.    

年代:1989

数据来源: WILEY

摘要:

AbstractA non‐parametric method for supervised pattern recognition is presented. The method is of the class modelling type, meaning that a classification rule is developed for each class, using the dissimilarities between the objects of the class. The dissimilarities between the objects within a class are related to the distances between all pairs of training objects. As distance metric, a measure is proposed that takes the correlation between the interval‐scale variables into account, and that moreover can be used for mixed types of variables. The classification rule is based on the construction of a boundary in the measurement space. For the determination of the class boundary, several strategies are proposed and compared.The performance of the technique is evaluated on the basis of several data sets. Comparison with the class modelling technique UNEQ shows its usefulness for practical applicat

ISSN:0886-9383    

DOI:10.1002/cem.1180030206

出版商:John Wiley&Sons, Ltd.    

年代:1989

数据来源: WILEY

摘要:

AbstractCalibrations to predict crude protein (CP) andin vitrodry matter digestibility (IVDMD) in dried grass silage from reflectance data collected at 19 wavelengths on an InfraAlyzer 400R have been developed using stepwise multiple linear (SML) and principal component (PC) regression techniques. A direct comparison of the efficacy of each multivariate technique in this applications has been possible by using identical calibration development and evaluation sample sets. The effect of two data transformation steps prior to PC regression was also investigated. PC regression of raw reflectance data yielded no significant improvement in the standard errors of prediction (SEP) for CP and IVDMD over those obtained by SMLR, viz. 0·61 vs 0·63 and 2·9 vs 3·0 respectively. Computation time for development and evaluation of the PC regression equation was less than for selection of the best SMLR equation, and PCR equations may be more robust. Data transformation to reduce granularity effects prior to PCR did not produce any improvement in predictive accuracy for either IVDMD or

ISSN:0886-9383    

DOI:10.1002/cem.1180030207

出版商:John Wiley&Sons, Ltd.    

年代:1989

数据来源: WILEY

摘要:

AbstractCare is required for multicomponent analysis if misleading results are to be avoided. The problem of ill‐conditioned calibration matrices is of primary concern. This type of numerical instability is represented as spectral overlap of calibration spectra. Depending on the degree of spectral overlap, the sample concentration estimates can be severely affected. A practical statistical procedure is discussed which tests for the presence of spectral overlap among the pure‐component spectra and simultaneously assesses the degree that concentration estimates may be degraded. Guidelines are developed to ascertain how much departure from spectral orthogonality is accepta

ISSN:0886-9383    

DOI:10.1002/cem.1180030208

出版商:John Wiley&Sons, Ltd.    

年代:1989

数据来源: WILEY

摘要:

AbstractImages can contain chemical information and many chemical methods can generate image data. For an efficient extraction of chemical data from images, data analysis techniques are necessary. It is a great advantage to be able to work on multivariate images. Many imaging techniques allow the extraction of chemical information. Inorganic analytical chemistry seems to have the longest tradition here, but organic chemistry and biochemistry may soon be catching up. Also large data arrays from non‐imaging techniques can be combined with image analysis in a useful way, provided certain conditions are fulfille

ISSN:0886-9383    

DOI:10.1002/cem.1180030209

出版商:John Wiley&Sons, Ltd.    

年代:1989

数据来源: WILEY

摘要:

AbstractWe consider the problem of resolving the total luminescence spectra of two‐component mixtures. Non‐negativity of the constituent spectra is commonly used as a criterion to accomplish this resolution, but its application seldom leads to unique answers. Sometimes, the condition must be relaxed because no solution satisfying non‐negativity exists. In these cases judgments must be made concerning which parts of the spectra will be allowed to go negative and by how much.The wavelength component vectorgramn (WCV) is presented here as a tool for assisting the experimenter in making these judgments. Even when non‐negative resolutions exist, it may happen that resolutions with slight negative elements will more closely match the spectra of the pure constituents. The WCV can also be used as a tool for detecting these cases. The methodology is illustrated with real data from a two‐componen

ISSN:0886-9383    

DOI:10.1002/cem.1180030210

出版商:John Wiley&Sons, Ltd.    

年代:1989

数据来源: WILEY

摘要:

AbstractConcentration estimates of components present in a sample mixture can be obtained using matrix mathematics. In the past, the condition number of the calibration matrix has been used to give an amplification factor by which uncertainties in data can work through to errors in the concentration estimates. This paper explores an additional interpretation of condition numbers with regards to significant figures and rounding errors. A procedure is suggested which will always give the most accurate concentration estimates provided the calibration matrix is not too ill‐conditioned. Condition numbers have also been used by analytical chemists to discuss the error bounds for concentration estimates. Unfortunately, only one representative error bound can be approximated for all the components. This paper will show how to compute bounds for individual concentration estimates obtained as solutions to a system ofmequations andnunknowns. The procedure is appropriate when calibration data and sample responses are inaccurat

ISSN:0886-9383    

DOI:10.1002/cem.1180030211

出版商:John Wiley&Sons, Ltd.    

年代:1989

数据来源: WILEY

ISSN:0886-9383    

DOI:10.1002/cem.1180030212

出版商:John Wiley&Sons, Ltd.    

年代:1989

数据来源: WILEY