Stacked interval sparse partial least squares regression analysis

Poerio, Dominic V.; Brown, Steven D.

doi:10.1016/j.chemolab.2017.03.006

Cited by 12 publications

(2 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The number of variables selected by using SPLS increases as more latent variables are used. When multiple components are required to model a signal, the SPLS solution is often saturated with variables . Selection of variables by using VIP can result in the selection of measured variables that have high variation but are not useful for prediction .…”

Section: Introductionmentioning

confidence: 99%

Using elastic net regression to perform spectrally relevant variable selection

Giglio

Brown

2018

Journal of Chemometrics

Self Cite

View full text Add to dashboard Cite

Multivariate data such as spectra frequently contain measured variables that are uninformative, and removal of such variables requires the use of methods that can be used to select informative variables. Partial least squares (PLS) regression may incorporate information from uninformative measured variables, and so it is important to select variables before performing the PLS regression. Elastic net (EN) regression can be used to perform variable selection automatically. An EN regression can be used to select groups of correlated variables or to select either sparse or nonsparse sets of variables. However, the predictive performance of the EN regression can be significantly worse than competing 1-step variable selection methods such as variable importance in projection (VIP). In the present work, the use of the EN to select variables, followed by conventional PLS regression on the selected variables (EN-PLS), has been investigated. Variable selection by using EN-PLS was compared with that from EN regression, sparse PLS regression, VIP, and from selectivity ratio selection on 2 data sets of visible/near-infrared spectra. In all cases, the wavelengths selected were compared with reference data. The variables selected by using EN-PLS offered advantages in interpretability and gave more robust prediction performance as compared with those obtained from full-spectrum PLS and the other variable selection methods. This paper reports a method for variable selection by using an EN regression prior to a second regression by using PLS, a 2-step method termed EN-PLS. Variables selected by using EN-PLS are compared with variables selected from the EN regression, as well as VIP, selectivity ratio, and the sparse PLS regression, 3 commonly used methods for variable selection in chemometrics. The EN-PLS is shown to select variables that were more easily interpreted. In addition, EN-PLS performed more robustly than a PLS regression performed on all variables, as well as reduced PLS regressions by using variables selected from either the sparse PLS regression algorithm or a VIP variable selection followed by PLS modeling.

show abstract

Section: Introductionmentioning

confidence: 99%

Using elastic net regression to perform spectrally relevant variable selection

Giglio

Brown

2018

Journal of Chemometrics

Self Cite

View full text Add to dashboard Cite

show abstract

“…Then the partial least square model of each combination is built, and the corresponding root mean square error of cross validation (RMSECV) is calculated [34][35][36]. The combination of subintervals with minimal RMSECV is optimal, and the model based on this optimal combination is considered to be the best [37].…”

Section: Introductionmentioning

confidence: 99%

Using Near Infrared Spectroscopy to Quickly Analyze Different Nitrogens during the Shortcut Biological Removal of Nitrogen from a Polluted River

Jian¹,

Li²,

Zhang³

et al. 2019

Pol. J. Environ. Stud.

View full text Add to dashboard Cite

During shortcut biological nitrogen removal in a polluted river, total nitrogen, ammonia nitrogen and nitrite nitrogen were quantified by near infrared spectroscopy and the synergy interval partial least squares (siPLS) algorithm. Spectral data of 138 water samples were obtained with a near infrared spectrometer. In addition, the real values of total nitrogen, ammonia nitrogen and nitrite nitrogen were measured with traditional chemical methods. SiPLS analysis models of total nitrogen, ammonia nitrogen and nitrite nitrogen were built through the siPLS algorithm based on spectral data and real values. The results obtained from the siPLS analysis model of total nitrogen revealed that, when the full spectra were divided into 19 intervals, the combination of the 7 th , 12 th and 19 th subintervals yielded the best model. The correction coefficient (R p ) is 0.9931, with the root mean squared error of calibration (RMSECV) being 1.7869. The results obtained from the siPLS analysis model of ammonia nitrogen indicated that, when the full spectra were divided into 16 intervals, the combination of the 1 st , 7 th , 15 th and 16 th subintervals yielded the best model. The R p is 0.9947 and the RMSECV is 1.3419. For nitrite nitrogen, the siPLS analysis model indicated that, when the full spectra were divided into 16 intervals, the combination of the 7th and the 11th subintervals yielded the best model. The R p and RMSECV was 0.9951 and 1.0518. These findings demonstrated that the proposed approach may effectively analyze the concentrations of total nitrogen, ammonia nitrogen and nitrite nitrogen during the treatment of a polluted river based on shortcut biological nitrogen removal. This approach, *

show abstract

Calibration Methodologies

Kalivas¹,

Brown²

2020

Comprehensive Chemometrics

View full text Add to dashboard Cite

Stacked interval sparse partial least squares regression analysis

Cited by 12 publications

References 27 publications

Using elastic net regression to perform spectrally relevant variable selection

Using elastic net regression to perform spectrally relevant variable selection

Using Near Infrared Spectroscopy to Quickly Analyze Different Nitrogens during the Shortcut Biological Removal of Nitrogen from a Polluted River

Calibration Methodologies

Contact Info

Product

Resources

About