RoBoost-PLS2-R: An extension of RoBoost-PLSR method for multi-response

Metz, Maxime; Ryckewaert, Maxime; Garcia, Sílvia Mas; Bendoula, Ryad; Dardenne, Pierre; Lesnoff, Matthieu; Roger, Jean‐Michel

doi:10.1016/j.chemolab.2022.104498

Cited by 6 publications

(32 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…One of the ideas behind using multiple criteria to update sample weights is to simply compute their product. However, in some recent works, 8,9 it is mentioned that the user can also try summation or averaging.…”

Section: Comments On the Algorithmmentioning

confidence: 99%

“…Hence, it is important to tune the α parameter, which can be performed using approaches such as cross-validation or using a separate validation set as performed in earlier studies. 9 The sample weights are updated as u i ( ð1 À u 2 i Þ 2 for ju i j ¼ < 1 and ju i j ( 0 for ju i j > 1.…”

Section: Comments On the Algorithmmentioning

confidence: 99%

“…Estimating the weights individually for each response was noted as more robust when the responses have different variances. 9 The irCovSel like the irPLS, PRM and RoBoost PLS approaches is a step-wise approach. This means that irCovSel selects one feature at a time, which gives it full freedom to have a wide range of extensions such as for selecting features in multiblock and multiway data.…”

Section: Comments On the Algorithmmentioning

confidence: 99%

“…Although major attention has been paid to the development of robust PLS methods [4][5][6][7][8][9] and a wide range of feature selection methods are available, [10][11][12][13] there is currently limited literature on selection of features in the presence of outliers, that is, robust feature selection. Typically, such methods use some sort of subsampling to estimate feature stability from an ensemble of models, [14][15][16][17] that is, quite computationally intensive.…”

mentioning

confidence: 99%

See 3 more Smart Citations

Iterative re‐weighted covariates selection for robust feature selection modelling in the presence of outliers (irCovSel)

Mishra

Liland

2022

Journal of Chemometrics

View full text Add to dashboard Cite

A new method for feature selective modelling in the presence of outliers is presented. The method is a combination of iterative re-weighted partial least squares and the covariates selection approach. The method relies on iterative down-weighting of the outlying samples prior to estimating the squared covariance for covariates selection. In this way, the outlying samples carrying low weights have minimal influence on the squared covariance estimation, while the inliers have the maximum influence on the squared covariance estimation.The method allows selecting robust features, and models based on such features in general perform better in terms of prediction accuracy (lower error) than selecting features using equal sample weights for all samples. The algorithm description and tests of the method in single and multiple response scenarios are presented. Method performance is also demonstrated on a real spectral data set.

show abstract

Section: Comments On the Algorithmmentioning

confidence: 99%

Section: Comments On the Algorithmmentioning

confidence: 99%

Section: Comments On the Algorithmmentioning

confidence: 99%

mentioning

confidence: 99%

See 2 more Smart Citations

Iterative re‐weighted covariates selection for robust feature selection modelling in the presence of outliers (irCovSel)

Mishra

Liland

2022

Journal of Chemometrics

View full text Add to dashboard Cite

show abstract

“…The PLSR model was created using PLS_Toolbox 31 (Eigenvector Research, inc., WA), which works in a Matlab environment. Among the predictive models to be built on spectra dataset [32][33][34][35][36][37][38][39] (considering clustering or non-clustering models), we focused on PLSR2 40 because we aimed at estimating semi-quantitative concentrations of CH2Cl2 and C6H12 in a two-variable outputs (M = 2). We used 12 new spectra as a test set to investigate the model prediction ability, measured on a single gas or a combination of the two.…”

Section: Spectra Analysismentioning

confidence: 99%

Towards Real-Time Portable Gas Mass Spectroscopy Detection by Multivariate Analysis: a Case Study

Franceschelli

Ciricugno

Lorenzo

et al. 2022

Preprint

View full text Add to dashboard Cite

Early and significant results for a real-time, column-free miniaturized gas mass spectrometer (MS) in detecting target species with partial overlapping spectra are reported. The achievements have been possible using both nanoscale holes to be used as a nanofluidic sampling inlet system and a robust statistical technique based on multivariate analysis to build predictive models. Even if the presented physical implementation could be used with gas chromatography (GC) columns, the aim of high miniaturization requires investigating its detection performance with no GC aid. For this reason, suitable analytical models were studied to get a semi-quantitative evaluation with very low computational resources. As a study case, dichloromethane (CH2Cl2) and cyclohexane (C6H12) with concentrations in the 6-93ppm range in single and compound mixtures were used. The nano-orifice approach was able to acquire raw spectra in 60 seconds with correlation coefficients of 0.525 and 0.578 with respect to the NIST reference database, respectively. Then, we built a calibration dataset on 2277 raw spectra of 10 known different mixtures using partial least square regression (PLSR) for statistical data inference. The model showed a normalized full-scale root-mean square deviation (NRMSD) accuracy of \(10.9\text{\%}\) and \(18.4\text{\%}\) for each species, respectively, even in combined mixtures.

show abstract

An algorithm for robust multiblock partial least squares predictive modelling

Mishra

Liland

2023

Journal of Chemometrics

View full text Add to dashboard Cite

A new algorithm for robust multiblock (data fusion) modelling in the presence of outlying observations is presented. The method is a combination of a robust modelling technique called iterative reweighted partial least squares and the block order and scale-independent component-wise multiblock partial least squares modelling. The method is based on automatic down-weighting of outlying observations such that their contribution is minimal during the estimation of block-wise partial least squares models, thus leading to robust modelling minimally affected by outliers. The algorithm and test of the methods for modelling multiblock data sets (simulated and real) in the presence of outlying observation are demonstrated.

show abstract

RoBoost-PLS2-R: An extension of RoBoost-PLSR method for multi-response

Cited by 6 publications

References 14 publications

Iterative re‐weighted covariates selection for robust feature selection modelling in the presence of outliers (irCovSel)

Iterative re‐weighted covariates selection for robust feature selection modelling in the presence of outliers (irCovSel)

Towards Real-Time Portable Gas Mass Spectroscopy Detection by Multivariate Analysis: a Case Study

An algorithm for robust multiblock partial least squares predictive modelling

Contact Info

Product

Resources

About