Multidimensional single-index signal regression

BD Marx, Paul Eilers, Baoyue Li

Research output: Contribution to journalArticleAcademicpeer-review

10 Citations (Scopus)

Abstract

In general, linearity is assumed to hold in multivariate calibration (MVC), but this may not be true. We approach the MVC problem using multidimensional penalized signal regression, which can be extended with an explicit link function between linear prediction and response and in the spirit of single-index models. As the two-dimensional surface of calibration coefficients is smoothly and generally estimated with tensor product P-splines, the unknown link function is estimated using univariate Psplines. The methods presented are grounded in penalized regression, where difference penalties are placed on the rows and columns of the tensor product coefficients, as well as on the link function coefficients, each having its own tuning parameter. An application to ternary mixture data shows that a non-linearity is present. Performance comparisons are made to standard penalized signal regression, not only demonstrating the nonlinear effect, but also improvements in external prediction. (C) 2011 Elsevier B.V. All rights reserved.
Original languageUndefined/Unknown
Pages (from-to)120-130
Number of pages11
JournalChemometrics & Intelligent Laboratory Systems
Volume109
Issue number2
DOIs
Publication statusPublished - 2011

Research programs

  • EMC NIHES-01-66-01

Cite this