Multi-core computation in chemometrics: case studies of voltammetric and NIR spectrometric analyses

Abstract

A aplicação de técnicas quimiométricas sofisticadas a grandes conjuntos de dados tem se tornado possível devido aos contínuos aprimoramentos tecnológicos em computadores comerciais. Recentemente, tais aprimoramentos têm sido obtidos principalmente através da introdução de processadores com múltiplos núcleos. Contudo, o uso eficiente de hardware com múltiplos núcleos requer o desenvolvimento de software apropriado para computação paralela. Este artigo trata da implementação de paralelismo empregando o Matlab Parallel Computing Toolbox, que requer somente pequenas modificações em códigos quimiométricos já existentes de modo a explorar os benefícios do processamento em múltiplos núcleos. Empregando essa ferramenta de software, mostra-se que implementações paralelas podem proporcionar expressivos ganhos computacionais. Em particular, considera-se o problema de seleção de variáveis empregando o algoritmo das projeções sucessivas e o algoritmo genético, bem como o uso de validação cruzada em mínimos quadrados parciais. Para ilustração, duas aplicações analíticas são apresentadas: determinação de proteína em trigo por espectrometria de reflectância no infravermelho próximo e classificação de óleos vegetais comestíveis por voltametria de onda quadrada. Empregando as implementações propostas para computação paralela, ganhos computacionais de até 204% foram obtidos.The application of sophisticated chemometrics techniques to large datasets has been made possible by continuing technological improvements in off-the-shelf computers. Recently, such improvements have been mainly achieved by the introduction of multi-core processors. However, the efficient use of multi-core hardware requires the development of software that properly address parallel computing. This paper is concerned with the implementation of parallelism using the Matlab Parallel Computing Toolbox, which requires only simple modifications to existing chemometrics code in order to exploit the benefits of multi-core processing. By using this software tool, it is shown that parallel implementations may provide substantial computational gains. In particular, the present study considers the problem of variable selection employing the successive projections algorithm and the genetic algorithm, as well as the use of cross-validation in partial least squares. For demonstration, two analytical applications are presented: determination of protein in wheat by near-infrared reflectance spectrometry and classification of edible vegetable oils by square-wave voltammetry. By using the proposed parallel computing implementations, computational gains of up to 204% were obtained.Keywords: parallel computation, successive projections algorithm, genetic algorithm, partial least squares, voltammetric analysis, near-infrared spectrometric analysis Soares et al. 1627 Vol. 21, No. 9, 2010 IntroductionModern techniques and instrumentation provide ever-growing amounts of data (in terms of variables and samples) that need to be processed for analytical purposes. Hyphenated methods 1 ...

Genetic Algorithm for Variable and Samples Selection in Multivariate Calibration Problems

“…The data set for multivariate calibration study is the same used by (Soares et al, 2010a), that consists of 755 visible near infrared spectra of whole-kernel wheat samples, which were initially used as shoot-out data in the 2008 International Diffuse Reflectance Conference (once http://www.idrcchambersburg.-org/shootout.html) and protein content is chosen as the property of interest. The following tests were performed to evaluate the proposed method GADC.…”

A GPU-Based Implementation of the Firefly Algorithm for Variable Selection in Multivariate Calibration Problems

“…The real dataset employed in this work consists of whole grain wheat samples, obtained from vegetal material from occidental Canadian producers. The standard data were determined at the Grain Research Laboratory as in work of [7] , [5] and [51] . The data set for the multivariate calibration study consists of 1090 Near-Infrared (NIR) spectra of whole-kernel wheat samples, which were used as shoot-out data in the 2008 International Diffuse Reflectance Conference ( http://www.idrc-chambersburg.org/shootout.html ).…”

Modern Metaheuristic with Multi-Objective Formulation for the Variable Selection Problem

“…The standard data were determined at the Grain Research Laboratory as in works of Paula et al (2014b) and Soares et al (2010;2013). The data set for the multivariate calibration study consists of 1090 NearInfrared (NIR) spectra of whole-kernel wheat samples, which were used as shoot-out data in the 2008 International Diffuse Reflectance Conference http://www.idrcchambersburg.org/shootout.html).…”