New probabilistic versions of the SIMCA and CLASSY Classification methods

Voet, Hilko van der; Coenegracht, Pierrme M.J.; Hemel, Jan

doi:10.1016/s0003-2670(00)85688-3

Cited by 18 publications

(7 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The terminology`inside model space' and`outside model space', first adopted by Van Der Voet et al [14], refers to the multidimensional space defined by a classification model that employs the k largest principal components (the inside model space) and the complementary space defined by the residuals of the k-factor principal component model (the outside model space). The outside model space bootstrapping technique randomly resamples residuals with replacement using the method of Faber [17].…”

Section: Methodsmentioning

confidence: 99%

“…Inexpensive and increased computing capability has led to the increased application of the bootstrap to chemical estimation problems in areas such as thermodynamics, kinetics and spectroscopy. For example, the bootstrap and a related technique known as the jackknife method have been successfully used to estimate confidence intervals for reaction rate constants [13,14]. At times when it is not possible to obtain sufficient data for testing and validation, the leaveone-out jackknife method can be used to estimate the standard errors in PCR and PLS regression coefficients [15].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Bootstrap methods for assessing the performance of near‐infrared pattern classification techniques

Smith

Gemperline

2002

Journal of Chemometrics

View full text Add to dashboard Cite

Two parametric bootstrap techniques were applied to near-infrared (NIR) pattern classification models for two classes of microcrystalline cellulose, Avicel 1 PH101 and PH102, which differ only in particle size. The development of pattern classification models for similar substances is difficult, since their characteristic clusters overlap. Bootstrapping was used to enlarge small test sets for a better approximation of the overlapping area of these nearly identical substances, consequently resulting in better estimates of misclassification rates. A bootstrap that resampled the residuals, referred to as the outside model space bootstrap in this paper, and a novel bootstrap that resampled principal component scores, referred to as the inside model space bootstrap, were studied. A comparison revealed that classification rates for both bootstrap techniques were similar to the original test set classification rates. The bootstrap method developed in this study, which resampled the principal component scores, was more effective for estimating misclassification volumes than the residual-resampling method.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Bootstrap methods for assessing the performance of near‐infrared pattern classification techniques

Smith

Gemperline

2002

Journal of Chemometrics

View full text Add to dashboard Cite

show abstract

“…It is not particularly designed to model data from several sources, but it aims to develop a hybrid model that is superior to the individual models. One of the fi rst hybrid methods, CLASSY (Van der Voet and Coenegracht, 1987 ), was developed in 1987. It tries to take advantage of combining SIMCA and ALLOC in a single technique.…”

Section: Data and Model Fusionmentioning

confidence: 99%

Chemometrics in studies of food origin

Vandeginste¹

2013

New Analytical Approaches for Verifying the Origin of Food

View full text Add to dashboard Cite

“…The so-called "normal range" defined by Wold is omitted here (73). To avoid the accumulation of rounding errors due to inaccuracies in the scores associated with very small eigenvalues, the alternate formula in eq 10 is actually used, where ay is the raw absorbance spectrum and ay* is the s,•2 = ( ;• -ay*')2/(pk) (10) 7=i reproduced absorbance spectrum from the abstract solution.…”

Section: =1mentioning

confidence: 99%

“…One interesting observation that was not previously recognized in the pattern recognition literature is the complementary nature of classification using the Mahalanobis dis-tance and SIMCA residual variance in the principal component pattern space. In order to explain the complementary behavior of these two classification methods, the "inside model space" (IMS) and "outside model space" (OMS) defined by Van Der Voet et al is used (10). During principal component analysis of a training set, a set of eigenvectors is produced that can be divided into two sets: a primary set of eigenvectors that span the "inside model space" of the training set and a secondary set of eigenvectors that span the "outside model space" of the training set.…”

Section: Introductionmentioning

confidence: 99%

Combination of the Mahalanobis distance and residual variance pattern recognition techniques for classification of near-infrared reflectance spectra

Shah¹,

Gemperline²

1990

Anal. Chem.

View full text Add to dashboard Cite

465proaches agree well, a consequence of the area parameter of eq 2 ( N w J dominating at short times (i.e., in the time range where eq 1 was applied). The values of ro and N in Table I11 characterize the random (site overlap allowed) disk electrode ensemble that would display transient diffusion current equivalent to the corresponding composite. It is both surprising and encouraging that the transient diffusion current at these composite electrodes, whose surface morphologies are essentially impossible to defiie geometrically, can be predicted from model ensembles consisting of well-defined (disk) active site geometry. CONCLUSIONKel-F/ precious metal composite electrodes with low volume percent conductor have been shown to possess high electrical conductivities and to exhibit behavior typical of microelectrode ensembles. The enhanced current densities displayed by these composites should result in higher signal-to-noise ratios than obtained at the corresponding macro disk electrodes, making them attractive alternative electrode materials for electroanalytical measurements, including amperometric detection in flowing streams (24). Furthermore, the enhanced current density coupled with the high conductivity and low precious metal content of these composites may make them attractive for energy generation and storage applications, particularly where weight and cost are considerations. Further optimization of the fabrication and composition of Kel-F/precious metal composites should lead to improvements in the advantageous properties reported here for these first generation Kelgold and Kelplat electrodes. Work in these areas is in progress. Principal component analysis of near-infrared reflectance (NIR) spectra is used for the calcuiatlon of Mahalanoblseilng Of 'lass ('IMCA) ciasslfication models. The Mahaianobls dlstance classlflcatlon and the space spanned by the secondary vectors 1 s used in SIMCA residual variance ciasslfkation. The complementary behavior of these two classlflcatlon methods Is discussed and a new classification rule based on a combination of the two methods IS described.The application of NIR spectroscopy and the pattern recognltion technique for ldentlfylng and classifying raw materials used in pharmaceutical Industry is also discussed.

show abstract

New probabilistic versions of the SIMCA and CLASSY Classification methods

Cited by 18 publications

References 11 publications

Bootstrap methods for assessing the performance of near‐infrared pattern classification techniques

Bootstrap methods for assessing the performance of near‐infrared pattern classification techniques

Chemometrics in studies of food origin

Combination of the Mahalanobis distance and residual variance pattern recognition techniques for classification of near-infrared reflectance spectra

Contact Info

Product

Resources

About