Abstract:Partial least squares (PLS) regression has achieved desirable performance for modeling the relationship between a set of dependent (response) variables with another set of independent (predictor) variables, especially when the sample size is small relative to the dimension of these variables. In each iteration, PLS finds two latent variables from a set of dependent and independent variables via maximizing the product of three factors: variances of the two latent variables as well as the square of the correlation between these two latent variables. In this paper, we derived the mathematical formulation of the relationship between mean square error (MSE) and these three factors. We find that MSE is not monotonous with the product of the three factors. However, the corresponding optimization problem is difficult to solve if we extract the optimal latent variables directly based on this relationship. To address these problems, a novel multilinear regression model-variance constrained partial least squares (VCPLS) is proposed. In the proposed VCPLS, we find the latent variables via maximizing the product of the variance of latent variable from dependent variables and the square of the correlation between the two latent variables, while constraining the variance of the latent variable from independent variables must be larger than a predetermined threshold. The corresponding optimization problem can be solved computational efficiently, and the latent variables extracted by VCPLS are near-optimal. Compared with classical PLS and it is variants, VCPLS can achieve lower prediction error in the sense of MSE. The experiments are conducted on three near-infrared spectroscopy (NIR) data sets. To demonstrate the applicability of our proposed VCPLS, we also conducted experiments on another data set, which has different characteristics from NIR data. Experimental results verified the superiority of our proposed VCPLS.

Locality-Preserving Partial Least Squares Regression

Locality Preserving Projections (LPP)

Locality Pursuit Embedding

Maximum Likelihood Estimators in a Two Step Model for PLS

Orthogonal Locality Preserving Indexing

Learning a Locality Preserving Subspace for Visual Recognition.

Optimal Weighting Distance-Based Similarity for Locally Weighted PLS Modeling

Partial least trimmed squares regression

Multilinear and Nonlinear Generalizations of Partial Least Squares: an Overview of Recent Advances

New Robust Projection to Latent Structure

Partial Least Squares Methods: Partial Least Squares Correlation and Partial Least Square Regression

Statistical and computational analysis of locality preserving projection

A Locality Preserving Approach for Kernel PCA

Deep Learning Partial Least Squares

Direct projection to latent variable space for fault detection

MLPP: A modified version of locality preserving projection

Nonlinear Projection to Latent Structures Method and Its Applications

[HDDA] sparse subspace constrained partial least squares

Variance constrained partial least squares

A non-asymptotic analysis of the single component PLS regression

Jointly Sparse Global SIMPLS Regression