Selecting the optimum number of partial least squares components for the calibration of attenuated total reflectance-mid-infrared spectra of undesigned kerosene samples

Gómez-Carracedo, M.P.; Andrade, J.M.; Rutledge, Douglas N.; Faber, Nicolaas M.

doi:10.1016/j.aca.2006.12.036

Cited by 52 publications

(24 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This criterion is based on q-fold CV methods (Breiman et al 1984) and was computed for both q = n, leading to the universal standard CV method called leave-one-out CV (Gómez-Carracedo et al 2007), and q = 5 (5-CV) following recommendations of Kohavi (1995) and Hastie et al (2009), so as to reduce variability in the CV method. The BIC criteria, corrected with the estimated degrees of freedom (DoF) by Krämer and Sugiyama (2011), was also included since to the best of our knowledge, no published study has analyzed its performance.…”

Section: Existing Criteria Used For Comparisonmentioning

confidence: 99%

“…However, it has also been established that though the use of the bootstrap for predictive error estimation efficiently reduces the variability issue, it can also lead to large bias (Efron and Tibshirani 1993, p. 255;Kohavi 1995). Much further literature is also available, introducing new criteria or comparing criteria: Höskuldsson (1996), Van der Voet (1994), Li et al (2002), Green and Kalivas (2002), Gourvenec et al (2003), and Gómez-Carracedo et al (2007) are some examples. Performing a global state-of-the-art review on this subject would be difficult due to the vast number of previous works.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A new universal resample-stable bootstrap-based stopping criterion for PLS component construction

et al. 2016

View full text Add to dashboard Cite

We develop a new robust stopping criterion for partial least squares regression (PLSR) component construction, characterized by a high level of stability. This new criterion is universal since it is suitable both for PLSR and extensions to generalized linear regression (PLSGLR). The criterion is based on a non-parametric bootstrap technique and must be computed algorithmically. It allows the testing of each successive component at a preset significance level α. In order to assess its performance and robustness with respect to various noise levels, we perform dataset simulations in which there is a preset and known number of components. These simulations are carried out for datasets characterized both by n > p, with n the number of subjects and p the number of covariates, as well as for n < p. We then use t-tests to compare the predictive performance of our approach with other common criteria. The stability property is in particular tested through re-sampling processes on a real allelotyping dataset. An important additional conclusion is that this new criterion gives globally better predictive performances than

show abstract

Section: Existing Criteria Used For Comparisonmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

A new universal resample-stable bootstrap-based stopping criterion for PLS component construction

et al. 2016

View full text Add to dashboard Cite

show abstract

“…Other recent papers give good overviews of the validation literature [57,67,70]. Wiklund et al [57] give a comprehensive introduction to the re-sampling situation in chemometrics.…”

Section: Cross-validation In Recent Chemometrics Historymentioning

confidence: 99%

“…Kohonen [69], in a recent thorough overview, remarks that full cross-validation (often termed LOOCV) is generally seen as the universal standard and refers to Gó mez-Carracedo et al [70] as the progenitor for this sweeping claim. Kohonen discusses the many historical uses of PRESS (Predicted Residual Sum-of-Squares).…”

Section: Cross-validation In Recent Chemometrics Historymentioning

confidence: 99%

Principles of Proper Validation: use and abuse of re‐sampling for validation

Esbensen¹,

Geladi

2010

Journal of Chemometrics

242

176

View full text Add to dashboard Cite

Validation in chemometrics is presented using the exemplar context of multivariate calibration/prediction. A phenomenological analysis of common validation practices in data analysis and chemometrics leads to formulation of a set of generic Principles of Proper Validation (PPV), which is based on a set of characterizing distinctions: (i) Validation cannot be understood by focusing on the methods of validation only; validation must be based on full knowledge of the underlying definitions, objectives, methods, effects and consequences-which are all outlined and discussed here. (ii) Analysis of proper validation objectives implies that there is one valid paradigm only: test set validation. (iii) Contrary to much contemporary chemometric practices (and validation myths), cross-validation is shown to be unjustified in the form of monolithic application of a one-for-all procedure (segmented cross-validation) on all data sets. Within its own design and scope, cross-validation is in reality a sub-optimal simulation of test set validation, crippled by a critical sampling variance omission, as it manifestly is based on one data set only (training data set). Other re-sampling validation methods are shown to suffer from the same deficiencies. The PPV are universal and can be applied to all situations in which the assessment of performance is desired: prediction-, classification-, time series forecasting-, modeling validation. The key element of PPV is the Theory of Sampling (TOS), which allow insight into all variance generating factors, especially the so-called incorrect sampling errors, which, if not properly eliminated, are responsible for a fatal inconstant sampling bias, for which no statistical correction is possible. In the light of TOS it is shown how a second data set (test set, validation set) is critically necessary for the inclusion of the sampling errors incurred in all 'future' situations in which the validated model must perform. Logically, therefore, all one data set re-sampling approaches for validation, especially cross-validation and leverage-corrected validation, should be terminated, or at the very least used only with full scientific understanding and disclosure of their detrimental variance omissions and consequences. Regarding PLS-regression, an emphatic call is made for stringent commitment to test set validation based on graphical inspection of pertinent t-u plots for optimal understanding of the X-Y interrelationships and for validation guidance. QSAR/QSAP forms a partial exemption from the present test set imperative with no generalization potential.

show abstract

“…Para optimizar el modelo se utilizaron varios métodos de estandarización, como son la centralización, el autoescalado y la transformación logarítmica [155].…”

Section: Análisis Estadístico De Los Datosunclassified

Desarrollo de un sistema de medida basado en espectroscopía de impedancia para la determinación de parámetros fisicoquímicos en alimentos

Peris¹

View full text Add to dashboard Cite

Si precisas una mano, recuerda que yo tengo dos San AgustínAna Fuentes ha hecho posible la validación del equipo. Ha sido la mano ejecutora en todos los trabajos de campo, llevando a cabo las determinaciones fisicoquímicas y las medidas de impedancia. Una trabajadora incansable, meticulosa y precisa, siempre dispuesta y muy profesional. Ella es una de las máximas responsables de que este trabajo vea la luz. No importaban las muestras ni las repeticiones, ella siempre ha convertido aquello que a mi me parecía inaccesible en sencillo. Muchas gracias por tu implicación y tu ayuda.Mi agradecimiento a Edgard por la realización de las determinaciones analíticas vinculadas al estudio de la evolución de la frescura de pescado. Un amigo es aquel que adivina siempre cuándo se le necesita Jules RenardA Ricardo Zahonero le quiero y admiro desde hace muchos años. Con él comencé mi andadura profesional y fue una suerte, porque su maestría me facilitó el rodaje y maquilló mi inexperiencia. Siempre he sabido que podía contar con él. En este trabajo ha sido el responsable de la fabricación de electrodos de medida del equipo implementado. Gracias amigo por tener siempre tiempo para mí, por tu talento y por tu cariño. Ayudadme a comprender lo que os digo y os lo explicaré mejor Antonio MachadoDe Jesús Brezmes valoro enormemente su capacidad para explicar con claridad los complejos conceptos del análisis multivariante. Gracias a la nitidez de su conferencia pude empezar a descifrar la esencia de las herramientas estadísticas aplicadas en esta tesis y su forma de transmitir conocimientos me dio fuerza y confianza para estudiarlas en profundidad.Juan Soto ha tenido la calma y paciencia necesarias para traducir a mi idioma conceptos de electroquímica. Muchas gracias por tu contribución a este trabajo.vii Los individuos marcan goles, pero los equipos ganan partidos Zig ZiglarEduardo García y Luís Gil me hicieron sentir desde el principio como un miembro más del equipo, sin condiciones, con la mano tendida. Soy muy afortunado por pertenecer a este grupo de trabajo donde no hay tensiones, vencedores, ni vencidos, donde se persigue un único objetivo, el de crecer todos juntos y permanecer unidos. Gracias por vuestra acogida. Ahora sé que juego en casa. El amor es lo único que crece cuando se reparte Antoine de Saint-ExuperySi hay alguien que merece reconocimiento y aplauso es mi madre. Una mujer con un corazón y un coraje enorme, una luchadora. Una persona completa y noble, dulce y tierna. Ella junto con Vicenta y Tona son las abanderadas de un gran matriarcado que han constituido la columna vertebral de mi infancia y adolescencia. Las tres sois unas mujeres excepcionales que me habéis hecho vivir muchos momentos felices. Es un orgullo poder teneros cerca. Gracias por vuestro amor.Me he criado en el entorno familiar muy numeroso, en el que convivimos aún hoy padres, tíos, primos y hermanos como una auténtica piña, algo difícil actualmente. Creo por tanto en la fuerza de la familia y de ella me he alimentado. A todos os agradezco el apoyo ...

show abstract

Selecting the optimum number of partial least squares components for the calibration of attenuated total reflectance-mid-infrared spectra of undesigned kerosene samples

Cited by 52 publications

References 22 publications

A new universal resample-stable bootstrap-based stopping criterion for PLS component construction

A new universal resample-stable bootstrap-based stopping criterion for PLS component construction

Principles of Proper Validation: use and abuse of re‐sampling for validation

Desarrollo de un sistema de medida basado en espectroscopía de impedancia para la determinación de parámetros fisicoquímicos en alimentos

Contact Info

Product

Resources

About