Harald Martens

Bio: Harald Martens is an academic researcher from Norwegian University of Science and Technology. The author has contributed to research in topics: Partial least squares regression & Regression analysis. The author has an hindex of 49, co-authored 146 publications receiving 11179 citations. Previous affiliations of Harald Martens include Norwegian Computing Center & Technical University of Denmark.

Linearization and Scatter-Correction for Near-Infrared Reflectance Spectra of Meat

Abstract: This paper is concerned with the quantitative analysis of multicomponent mixtures by diffuse reflectance spectroscopy. Near-infrared reflectance (NIRR) measurements are related to chemical composition but in a nonlinear way, and light scatter distorts the data. Various response linearizations of reflectance (R) are compared (R with Saunderson correction for internal reflectance, log 1/R, and Kubelka-Munk transformations and its inverse). A multi-wavelength concept for optical correction (Multiplicative Scatter Correction, MSC) is proposed for separating the chemical light absorption from the physical light scatter. Partial Least Squares (PLS) regression is used as the multivariate linear calibration method for predicting fat in meat from linearized and scatter-corrected NIRR data over a broad concentration range. All the response linearization methods improved fat prediction when used with the MSC; corrected log 1/R and inverse Kubelka-Munk transformations yielded the best results. The MSC provided simpler calibration models with good correspondence to the expected physical model of meat. The scatter coefficients obtained from the MSC correlated with fat content, indicating that fat affects the NIRR of meat with an additive absorption component and a multiplicative scatter component.

Multivariate Analysis of Quality : An Introduction

Abstract: The book provides an introduction to multivariate data analysis using linear modelling and its applications to quality assessment with special emphasis on chemometrics and sensory science. The aim of the book is to help students and researchers (the problem and data `owners') to analyse their large empirical data sets with minimum prior knowledge of algebra and mathematical statistics. The text (445 pages) is divided into four parts: overview (76 pp), methodology (156 pp), applications (122 pp) and appendices (72 pp). In the first part, the authors motivate a reader to use multivariate methods, explain concepts of quality assessment and provide `a layman's guide to multivariate data analysis'. The importance of background knowledge about the problem and good quality of input data is emphasized. A research project with the aim of obtaining new facts from an empirical data set is divided into six steps from the original question to the unfolded answer. These steps are then formally followed in all the presented examples. Finally, the principles of soft bilinear modelling (the basic data-analytic tool used in the book) are described at a glance. In the second part, the method of partial least squares regression is derived from the methods of linear least squares regression and principal component analysis, including its individual variants and applications (multivariate calibration, prediction, discrimination and classification). Individual chapters are dedicated to validation of results and experimental planning. Part three describes five specific experiments: analysis of NIR spectra, analysis of questionnaire data on the quality of the working environment, prediction of toxicity from chemical structure, quality monitoring of a sugar production process, and exploratory search for optimal conditions preventing loss of quality in stored food. Appendices in the final part provide additional information to every individual chapter. The text is supported by 97 figures, 19 tables and 114 references. Most derivations and explanations in the book are based on examples which are described in almost all possible details except calculations: the main emphasis is on experimental design, organization of input data tables and, above all, on interpretation and validation of results. Computing itself is supposed to be performed by some available data analytic software such as, e.g., The Unscrambler (http://www.camo.no) or PLS_Toolbox in Matlab http://www.mathworks.com). However, the book provides a rather general guide independent of any specific software. The beginner will probably benefit most from the great experience of the authors: the description of advantages and risks in multivariate data analysis is well and proportionally balanced and documented. Minimum abstraction and mathematical formalism may also bring the book closer to a wider readership. At the same time, the authors' approach is fair and responsible: they encourage readers to work independently but they clearly mark the limits beyond which the reader should seek help from a professional statistician. The absence of abstraction also does not necessarily prevent some generalization: while the reader will probably solve new problems by analogy with those described in the book, the examples presented are accompanied by lists of related problems in order to demonstrate the wide spectrum of possible applications. The strengths of the book determine also some of its limitations: although it is intended for `laymen' and contains minimal mathematics, it is not too easy to read. Many references to other parts of the book, descriptions of the authors' intentions and directions on how to read the book (often useful) sometimes distract rather than focus the reader's attention on the main problem. A diligent reader, however, will find the book useful: beginners and data-analytic software users as a good start and introduction, more advanced users and teachers as a possible source of inspiration and a good teaching aid. Martin Samal

Light scattering and light absorbance separated by extended multiplicative signal correction. application to near-infrared transmission analysis of powder mixtures.

Abstract: The extended multiplicative signal correction (EMSC) preprocessing method allows a separation of physical light-scattering effects from chemical (vibrational) light absorbance effects in spectra from, for example, powders or turbid solutions. It is here applied to diffuse near infrared transmission (NIT) spectra of mixtures of wheat gluten (protein) and starch (carbohydrate) powders, linearized by conventional log(1/T). Without any correction for uncontrolled light scattering variation between the powder samples, these absorbance spectra could give reasonable predictions of the analyte (gluten), but only when using multivariate calibration with a much more complex model than expected. Standard MSC preprocessing did not work for these data at all; it removed too much analyte information. However, the EMSC preprocessing yielded powder spectra that obeyed Beer's Law more or less as if they had been obtained from transparent liquid solutions, apparently by isolating the chemical light absorption from additive, multiplicative, and wavelength-dependent effects of uncontrolled light-scattering variations. The model-based EMSC and its converse, the extended inverted signal correction (EISC), gave rather complete descriptions of the diffuse absorbance spectra and virtually indistinguishable performance in the calibration set and the test set of samples.

The Design and Analysis of Experiments

Abstract: THE DESIGN AND ANALYSIS OF EXPERIMENTS. By Oscar Kempthorne. New York, John Wiley and Sons, Inc., 1952. 631 pp. $8.50. This book by a teacher of statistics (as well as a consultant for \"experimenters\") is a comprehensive study of the philosophical background for the statistical design of experiment. It is necessary to have some facility with algebraic notation and manipulation to be able to use the volume intelligently. The problems are presented from the theoretical point of view, without such practical examples as would be helpful for those not acquainted with mathematics. The mathematical justification for the techniques is given. As a somewhat advanced treatment of the design and analysis of experiments, this volume will be interesting and helpful for many who approach statistics theoretically as well as practically. With emphasis on the \"why,\" and with description given broadly, the author relates the subject matter to the general theory of statistics and to the general problem of experimental inference. MARGARET J. ROBERTSON

A National Customer Satisfaction Barometer: The Swedish Experience:

Abstract: Many individual companies and some industries monitor customer satisfaction on a continual basis, but Sweden is the first country to do so on a national level. The annual Customer Satisfaction Baro...