Abstract:Partial least square discriminant analysis (PLS-DA) has achieved a huge success in many research systems, such as quality evaluation of agricultural products and drug analysis. However, it can be found that PLS-DA has various forms which can be mainly divided into two strategies separately called as PLS1-DA and PLS2-DA which is overused. In this work, we proposed a novel strategy not only to select representative samples for modeling, but also to build qualitative models, which is called as principal component analysis Euclidean distance PLS1-DA (PCA-EuD-PLS1-DA). Outliers were detected by the method Leverage. It is very interesting to compare the results of traditional PLS-DA tactics with the proposed method on six datasets. EuD-PLS1-DA with PCA partitioning subsets has the optimal performances than both PLS2-DA and assigned PLS1-DA when confronting multi-class problems. In particular, different classes’ tablets in NIR can be well discriminated and different classes’ tablets in Raman can be discriminated in all EuD-PLS1-DA models; and their prediction accuracy is above 98% and 80%, respectively. Their prediction accuracy for PLS2-DA is above 56% and 26%, respectively. And all RMSECV in EuD-PLS1-DA is smaller than the remaining classifiers; so does RMSEP. This behavior means EuD-PLS1-DA has a better fitness than the remaining methods. When there are only two classes in a dataset, performance of all PLS-DA models is highly similar. Before comparing the results of PLS-DA derivatives, the shortcomings in the other two methods have been systematically described. Sample selection algorithms including K–S, SPXY and PCA were compared with percentage of chosen samples in every class. And the current ratios are mainly in the area of 70–90%. Besides, ratios of every class’ samples selected for modeling are highly similar among K–S, SPXY and PCA. Most importantly, the proposed method can be realized in most statistical software.

The Equivalence of Partial Least Squares and Principal Component Regression in the Sufficient Dimension Reduction Framework

Model Selection for Partial Least Squares Based Dimension Reduction

A Principal Square Response Forward Regression Method for Dimension Reduction

Subsampling for Partial Least-Squares Regression Via an Influence Function

Sufficient Dimension Reduction for Classification Using Principal Optimal Transport Direction

Author Response for "partial Least Median of Squares Regression"

Feature Selection For Partial Least Square Based Dimension Reduction

Joint Principal Component and Discriminant Analysis for Dimensionality Reduction

Sparse Unsupervised Dimensionality Reduction Algorithms

Supervised Linear Dimension-Reduction Methods: Review, Extensions, and Comparisons

On partial sufficient dimension reduction with applications to partially linear multi-index models

Advance of the Sufficient Dimension Reduction

On The Number Of Partial Least Squares Components In Dimension Reduction For Tumor Classification

Principal component analysis balancing prediction and approximation accuracy for spatial data

Y-SPCR: A New Dimensionality Reduction Method for Gene Expression Data Classification

On sufficient dimension reduction via principal asymmetric least squares

Asymptotic results for nonparametric regression estimators after sufficient dimension reduction estimation

A selective review of sufficient dimension reduction for multivariate response regression

Principal Fitted Components for Dimension Reduction in Regression

Learning Heterogeneity in Causal Inference Using Sufficient Dimension Reduction

Comparison of a Novel PLS1-DA, Traditional PLS2-DA and Assigned PLS1-DA for Classification by Molecular Spectroscopy