Improved variable reduction in partial least squares modelling based on Predictive-Property-Ranked Variables and adaptation of partial least squares complexity

Andries, Jan P.M.; Heyden, Yvan Vander; Buydens, L.M.C.

doi:10.1016/j.aca.2011.06.037

Cited by 51 publications

(23 citation statements)

References 55 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For preventing overfitting in PLSR, as well as in LW, KNN‐LW, and KNN‐L for each h , k × h , and k , respectively, the number of components n comp was firstly selected using a heuristic criterion. This criterion was R = 1 − r ( a + 1)/ r ( a ), where r is the error rate (RMSECV or ERRCV) and a the PLS model dimension. R represents the relative gain in efficiency after a new dimension is added in the model.…”

Section: Methodsmentioning

confidence: 99%

Comparison of locally weighted PLS strategies for regression and discrimination on agronomic NIR data

Lesnoff

Metz

Roger

2020

Journal of Chemometrics

View full text Add to dashboard Cite

In multivariate calibrations, locally weighted partial least squared regression (LWPLSR) is an efficient prediction method when heterogeneity of data generates nonlinear relations (curvatures and clustering) between the response and the explicative variables. This is frequent in agronomic data sets that gather materials of different natures or origins. LWPLSR is a particular case of weighted PLSR (WPLSR; ie, a statistical weight different from the standard 1/n is given to each of the n calibration observations for calculating the PLS scores/loadings and the predictions). In LWPLSR, the weights depend from the dissimilarity (which has to be defined and calculated) to the new observation to predict. This article compares two strategies of LWPLSR: (a) "LW": the usual strategy where, for each new observation to predict, a WPLSR is applied to the n calibration observations (ie, entire calibration set) vs (b) "KNN-LW": a number of k nearest neighbors to the observation to predict are preliminary selected in the training set and WPLSR is applied only to this selected KNN set. On three illustrating agronomic data sets (quantitative and discrimination predictions), both strategies overpassed the standard PLSR. LW and KNN-LW had close prediction performances, but KNN-LW was much faster in computation time. KNN-LW strategy is therefore recommended for large data sets. The article also presents a new algorithm for WPLSR, on the basis of the "improved kernel #1" algorithm, which is competitor and in general faster to the already published weighted PLS nonlinear iterative partial least squares (NIPALS). K E Y W O R D Sdiscrimination, locally weighted calibration, near-infrared spectroscopy, partial least squares, regression

show abstract

Section: Methodsmentioning

confidence: 99%

Comparison of locally weighted PLS strategies for regression and discrimination on agronomic NIR data

Lesnoff

Metz

Roger

2020

Journal of Chemometrics

View full text Add to dashboard Cite

show abstract

“…This RMSEV is followed by Expression with L = I with RMSEV = 0.0156 and worst is the RR model at RMSEV = 0.0227 (both model vectors are in Figure (a)). The sparse model vectors shown in Figure select from the same wavelength regions as identified by other sparse modeling approaches .…”

Section: Tikhonov Regularization and Some Variantsmentioning

confidence: 99%

Overview of two‐norm (L₂) and one‐norm (L₁) Tikhonov regularization variants for full wavelength or sparse spectral multivariate calibration models or maintenance

Kalivas

2012

Journal of Chemometrics

View full text Add to dashboard Cite

Building a multivariate calibration model is typically accomplished using partial least squares, principal component regression, or ridge regression, also derived as the standard form of Tikhonov regularization (TR). These approaches can be used in a full variable mode (full wavelengths for spectroscopic data) or with wavelength selection (bands and/or individual for sparse models). Calibration maintenance is an important aspect of multivariate calibration and describes the situation of maintaining acceptable predictions from a model over time. In terms of TR, this amounts to updating an existing model (determined under primary conditions) to handle new secondary conditions such as a new instrument, sample matrix, or environmental conditions. This paper overviews TR in its ability to form a primary calibration model or update a primary model to new secondary conditions while using full wavelengths or simultaneously selecting wavelengths to form respective sparse models. These objectives can be accomplished in two‐norm (L2) or one‐norm (L1) or combined formats. Also included is a TR design that minimizes the effect of the standardization set composition, i.e., reduces the effect of an outlier. Ongoing work that removes all reference samples from the TR framework is also described. Copyright © 2012 John Wiley & Sons, Ltd.

show abstract

“…The spectral noise is introduced by sensor limitations and particle-size differences [27]. Hence, improving the PLS method for statistical analysis of spectral data has become a main research focus [28][29][30]. Multivariate stepwise linear regression (MSLR) finds and selects the variables exerting the most significant influences on the dependent variables and outperforms ordinary meta-regression; however, this method cannot remove the multi-collinearity between independent variables.…”

Section: Introductionmentioning

confidence: 99%

The Application of Discrete Wavelet Transform with Improved Partial Least-Squares Method for the Estimation of Soil Properties with Visible and Near-Infrared Spectral Data

Wang

Fang

et al. 2018

Remote Sensing

View full text Add to dashboard Cite

This study evaluated whether wavelet functions (Bior1.3, Bior2.4, Db4, Db8, Haar, Sym4, and Sym8) and decomposition levels (Levels 3-8) can estimate soil properties. The analysis is based on the discrete wavelet transform with partial least-squares (DWT-PLS) method, incorporated into a visible and near-infrared reflectance analysis. The improved DWT-PLS method (called DWT-Stepwise-PLS) enhances the accuracy of the quantitative analysis model with DWT-PLS. The cation exchange capacity (CEC) was best estimated by the DWT-PLS model using the Haar wavelet function. This model yielded the highest coefficient of determination (R v 2 = 0.787, p < 0.001), with the highest relative percentage deviation (RPD = 2.047) and lowest root mean square error (RMSE = 4.16) for the validation data set of the CEC. The RPD of the SOM predictions by DWT-PLS using the Bior1.3 wavelet function was maximized at 1.441 (R v 2 = 0.642, RMSE = 5.96), highlighting the poor overall predictive ability of soil organic matter (SOM) by DWT-PLS. Furthermore, the best performing decomposition levels of the wavelet function were distributed in the fifth, sixth, and seventh levels. For various wavelet functions and decomposition levels, the DWT-Stepwise-PLS method more accurately predicted the quantified soil properties than the DWT-PLS model. DWT-Stepwise-PLS using the Haar wavelet function remained the best choice for quantifying the CEC (R v 2 = 0.92, p < 0.001, RMSE = 4.91, and RPD = 3.57), but the SOM was better predicted by DWT-Stepwise-PLS using the Bior2.4 wavelet function (R v 2 = 0.8, RMSE = 5.34, and RPD = 2.24) instead of the Bior1.3 wavelet function. However, the performance of the DWT-Stepwise-PLS method tended to degrade at high and low decomposition levels of the DWT. These degradations were attributed to a lack of sufficient information and noise, respectively.

show abstract

Improved variable reduction in partial least squares modelling based on Predictive-Property-Ranked Variables and adaptation of partial least squares complexity

Cited by 51 publications

References 55 publications

Comparison of locally weighted PLS strategies for regression and discrimination on agronomic NIR data

Comparison of locally weighted PLS strategies for regression and discrimination on agronomic NIR data

Overview of two‐norm (L₂) and one‐norm (L₁) Tikhonov regularization variants for full wavelength or sparse spectral multivariate calibration models or maintenance

The Application of Discrete Wavelet Transform with Improved Partial Least-Squares Method for the Estimation of Soil Properties with Visible and Near-Infrared Spectral Data

Contact Info

Product

Resources

About

Improved variable reduction in partial least squares modelling based on Predictive-Property-Ranked Variables and adaptation of partial least squares complexity

Cited by 51 publications

References 55 publications

Comparison of locally weighted PLS strategies for regression and discrimination on agronomic NIR data

Comparison of locally weighted PLS strategies for regression and discrimination on agronomic NIR data

Overview of two‐norm (L2) and one‐norm (L1) Tikhonov regularization variants for full wavelength or sparse spectral multivariate calibration models or maintenance

The Application of Discrete Wavelet Transform with Improved Partial Least-Squares Method for the Estimation of Soil Properties with Visible and Near-Infrared Spectral Data

Contact Info

Product

Resources

About

Overview of two‐norm (L₂) and one‐norm (L₁) Tikhonov regularization variants for full wavelength or sparse spectral multivariate calibration models or maintenance