Tormod Næs

• PhD
• NOFIMA Norwegian Institute of Food, Fisheries and Aquaculture Research

In various situations requiring empirical model building from highly multivariate measurements, modelling based on partial least squares regression (PLSR) may often provide efficient low‐dimensional model solutions. In unsupervised situations, the same may be true for principal component analysis (PCA). In both cases, however, it is also of interes...

Cluster analysis is often used to group consumers based on their hedonic responses to products. We give a motivating example in which conventional cluster analyses converge on a solution where consumers do not agree on which products they like. We show why this occurs. We state a goal: to group together consumers who have a shared opinion of which...

Different approaches for handling consumer segments in L‐shape data are compared in a study conducted in Norway. Consumers evaluated eight different yoghurt samples with profiles varying in three intrinsic attributes following a full factorial design. Three blocks of data were collected including sensory properties, liking ratings, and consumer att...

Principal component analysis (PCA) is often used to explore sensory and consumer test data about products on multicollinear sensory attributes. In this paper, we propose an approach for investigating paired comparisons between products and their uncertainties in the principal components. We use the truncated total bootstrap (TTB) procedure to simul...

In sensory data analysis, assessor-dependent scaling effects may hinder the analysis of product differences. Romano et al. (2008) compared several approaches to reduce scaling differences between assessors by their ability to maximise the product effect F-values in a mixed ANOVA analysis. Their study on a sensory dataset of 14 cheese samples assess...

Rapid sensory profiling of a product over time presents a challenge in quality control, particularly when the product has a complex aroma and flavour profile. Polarized sensory positioning (PSP), based on the comparison of samples to fixed references, allows for rapid sensory profiling and data collection across multiple sessions. The efficacy of p...

External preference mapping is widely used in marketing and R&D divisions to understand the consumer behaviour. The most common preference map is obtained through a two-step procedure that combines principal component analysis and least squares regression. The standard approach exploits classical regression and therefore focuses on the conditional...

This chapter discusses a number of alternative optimisation criteria and also methods for situations where the samples are split into subgroups. Based on PCovR several extensions of multiblock methods are developed: sparse multiblock principal covariates regression and multiblock multiway covariates regression. The PCovR framework allows for extens...

This chapter describes the general structures of multiblock data sets. It is difficult to give a true taxonomy because a systematic classification of all the different multiblock methods is impossible to give. The chapter aims to structure the methods according to some characteristics. It uses the terminology from graph theory. In that context, a m...

This chapter discusses methods that incorporate a special kind of information, namely if the samples are collected according to an experimental design. A class of methods that can deal with (multiblock) high‐dimensional data with an underlying experimental design is ANOVA‐simultaneous component analysis (ASCA) which combines analysis of variance te...

A much used metric in data analysis to express the relationship between variables is the correlation coefficient. Such a coefficient measures the linear relation between two variables (Pearson correlation) or its non‐metric version which measures the relationship between the ranks of the variables (Spearman correlation). Such measures exist indeed...

This chapter discusses the most commonly used unsupervised methods with simple undirected linking structures. It also discusses a method that can deal with multiway arrays and a method that can handle blocks containing measurements of different scale‐types. Given the complexity of the data and types of systems studied, there are many unsupervised m...

This chapter briefly presents some unsupervised methods. It aims to distinguish between methods focusing on common variation and methods that separate the different types of variation (common, local, distinct). Moreover, methods are available for a shared variable mode, a shared sample mode and methods for the case that both modes are shared. The c...

This chapter focuses on supervised multiblock data analysis. A typical situation where supervised methods are important is process modelling where measurements are made at different places and stages and the goal is to predict the end‐product properties and to understand how the different input data influence the output. One of the most well‐known...

This chapter focuses on data blocks that are related in more complex ways. The most important among the skeletons here is the L‐structure, but it briefly discusses the so‐called domino structure. The chapter also discusses sequential methods that handle either the shared sample mode (horizontal) or the shared variable mode (vertical) first and meth...

This chapter describes some basic models such as principal component analysis, principal component regression, partial least squares (PLS), simultaneous component analysis, and generalised canonical analysis that can be considered building blocks for many of the multiblock data analysis methods to follow. It also discusses general theories regardin...

A large proportion of the multiblock methods are available as software implementations both in open source software and through commercial software packages. This chapter describes the R package multiblock. This package includes methods that are unique to the package and imports a large range of methods from other packages. Overviews of basic packa...

Temporal check-all-that-apply (TCATA) data can be summarized and explored using principal component analysis (PCA). Here we analyze TCATA data on Syrah wines obtained from a trained sensory panel. We evaluate new and existing methods to explore the uncertainty in the PCA scores. To do so, we use the bootstrap procedure to obtain many virtual panels...

Consumers can be clustered based on their product-related check-all-that-apply (CATA) responses. We identify two paradoxes that can occur if these clusters are derived from conventional similarity coefficients. The first paradox is that clustering similar consumers can nullify within-cluster sensory differentiation of products. The second paradox i...

ConsumerCheck is an open source data analysis software tailored for analysis of sensory and consumer data. Since some of the implemented methods are generic, such as PCA, PLSR and PCR, other data from other domains may also be analysed with ConsumerCheck. The software comes with a graphical user interface and as such provides non-statisticians and...

Multiblock analysis attacks the problem of how to combine data from various data sources for purposes such as prediction, classification, clustering, or visual data analysis. A key concept is the distinction between “common” and “distinct” parts, that is, what information repeats itself across the blocks and what is unique to an individual block. T...

Different approaches for handling L-shaped data are compared for the first time in a study conducted with Norwegian consumers. Consumers (n = 101) valuated eight different yoghurt profiles varying in three intrinsic attributes such as viscosity, particle size, and flavour intensity following a full factorial design. Sensory attributes, consumers’ l...

Problems concerning covariance among independent variables are well understood and dealt with by inverse regression methods like partial least squares regression. However, covariance between dependent variables has only received minor attention. Biological samples are often complex mixtures of multiple covarying compounds. During multivariate calib...

This article presents a discussion of principal components analysis of descriptive sensory data. Focus is on standardization, many correlated variables, validation, and the use of descriptive data in preference mapping. Different ways of performing the analysis are presented and discussed with focus on how to obtain informative and reliable results...

Labels have been reported to influence children's perception and choice of foods. However, the influence of label information on parents’ food choices for their children has not received as much attention in the literature yet. In this context, the objectives of the present study were: i) to evaluate the impact of label information on parents' heal...

Quantitative Descriptive Analysis (QDA) was compared with novel sensory methodology on five high‐end loudspeakers using five tracks. The focus of the study was on the ability of different sensory methods to predict quality assessments (liking) made by consumers. Preference mapping indicates that projective mapping (PM) and Check‐All‐That‐Apply (CAT...

Due to its simplicity, Check-all-that-apply (CATA) is a promising method for consumer studies with children to generate sensory and other descriptions of samples, and to find their drivers of liking. This paper explores how children’s approach to the CATA test influences the outcome, based on two case studies that illustrate suitable setups for CAT...

This paper is about the use of the multiblock regression method sequential and orthogonalized partial least squares (SO‐PLS) for path modeling. The paper is a follow up of previously published papers on the same topic and presents a number of new results for the method. First of all, the paper discusses more thoroughly the aspect of how to incorpor...

In an investigation of objective measures that link grape composition to wine quality, this study sought to identify Cabernet Sauvignon grape parameters that predict the sensory properties of the corresponding wines. Eleven chemical measures comprising volatile and non-volatile compounds, enzyme activity plus standard industry harvest measurements...

In the development of sensory and consumer science, data are often collected in several blocks responding to different aspects of consumer experience. Sometimes the task of organizing the data and explaining their relation is non-trivial, especially when considering structural (casual) relationship between data sets. In this sense, PLS path modelli...

Temporal Dominance of Sensations (TDS) and Temporal Check-all-that-Apply (TCATA) from three different case studies are compared by means of canonical correlation analysis, orthogonalization and principal component analysis of the vertically unfolded data (which means that the matrices compared have samples*timepoints in the rows and attributes in t...

Recent research suggests that a robot's motors make sounds that can influence users' perception of the robot's characteristics. To more deeply understand users' associations with specific sonic characteristics, we adapted methods from sensory science including Check All That Apply (CATA) questions and Polarized Sensory Positioning (PSP) to tease ou...

This paper presents a new approach to path modeling named SO‐PLS path modeling (SO‐PLS‐PM) and compares it with the more well‐known PLS path modeling (PLS‐PM). The new method is flexible and graphically oriented and allows for handling multidimensional blocks and diagnosing missing paths. In order to allow for a thorough comparison between the two...

In this chapter, the sequentially orthogonalized-PLS (SO-PLS) method and some of its main extensions are described and illustrated. Both theoretical aspects and applications on real data are discussed. SO-PLS is a multiblock regression method in which the information is extracted sequentially from the predictor blocks and there is no limitation in...

Expectations of satiation and satiety have been increasingly investigated because of the interest in how they, along with liking, can modulate portion-size selection. Consumer characteristics can also be important when consumers select their portion size. However, the contribution and interaction of consumers and product aspects to portion size sel...

consumers with difference in mouth behaviours

Different strategies for investigating individual differences among consumers using choice experiments are compared. The paper is based on a consumer study of iced coffee in Norway. Consumers (n = 102) performed a choice task of twenty different iced coffee profiles varying in coffee type, production origin, calorie content and price following an o...

Application of different multivariate measurement technologies to the same set of samples is an interesting challenge in many fields of applied data analysis. Our proposal is a 2‐stage similarity index framework for comparing 2 matrices in this type of situation. The first step is to identify factors (and associated subspaces) of the matrices by me...

Projective mapping (PM) or napping® has attained much attention in recent literature as a method for fast sensory profiling and measurement of consumer perception. However, little work has been done to understand the consumer's individual differences in these experiments. In this work, segmentation criteria based on the Procrustes distance are expl...

Online communities can be an attractive source of ideas for product and process innovations. However, innovative user-contributed ideas may be few. From a perspective of harnessing “big data” for inbound open innovation, the detection of good ideas in online communities is a problem of detecting rare events. Recent advances in text analytics and ma...

Freshwater scarcity is becoming one of the most pressing issues of the global environmental sustainability, and agriculture is the main responsible of that scarcity. During the last decade, there has been an increasing consumers’ environmental concern about the impact of food production on water usage. This paper investigates young consumers’ prefe...

The current study determined the applicability of sequential and orthogonalised-partial least squares (SO-PLS) regression to relate Cabernet Sauvignon grape chemical composition to the sensory perception of the corresponding wines. Grape samples (n = 25) were harvested at a similar maturity and vinified identically in 2013. Twelve measures using va...

Individual Differences in Sensory and Consumer Science: Experimentation, Analysis and Interpretation presents easily readable, state-of-the-art coverage on how to plan and execute experiments that give rise to individual differences, also providing the framework for successful analysis and interpretation of results. The book highlights the differen...

In analysis of variance simultaneous component analysis, permutation testing is the standard way of assessing uncertainty of effect level estimates. This article introduces an analytical solution to the assessment of uncertainty through classical multivariate regression theory. We visualize the uncertainty as ellipsoids, contrasting these to data e...

For describing the evolution of sensory properties during eating, dynamic sensory methods are still being developed and optimised. Temporal Dominance of Sensations (TDS) and Temporal Check All That Apply (TCATA) are currently the most used and discussed. The aim of this study was to compare TDS, TCATA and a variant of TDS, performed by modality (M-...

In conjoint analysis (CA), one generates a set of product profiles that vary systematically in terms of their attributes and attribute levels. These product profiles are then presented to consumers, who will state their preferences in a rating, ranking, or choice task. After presenting the conceptual background of CA methods, this chapter introduce...

Selecting the correct number of factors in principal component analysis (PCA) is a critical step to achieve a reasonable data modelling, where the optimal strategy strictly depends on the objective PCA is applied for. In the last decades, much work has been devoted to methods like Kaiser's eigenvalue greater than 1 rule, Velicer's minimum average p...

TDS describes the evolution of the dominant sensory attributes during consumption. Dominance can be assessed as the sensation that captures the attention, the most striking, or the new sensation that pops up, but not necessarily the most intense. This wide definition implies that individual assessors within a panel might assess dominance differentl...

Consumers in industrialized countries are nowadays much more interested in information about the production methods and components of the food products that they eat, than they had been 50 years ago. Some production methods are perceived as less “natural” (i.e. conventional agriculture) while some food components are seen as “unhealthy” and “unfami...

Different strategies for investigating individual differences among consumers using choice experiments are compared. The paper is based on a consumer study of iced coffee in Norway. Consumers (n = 102) performed a choice task of twenty different iced coffee profiles varying in coffee type, production origin, calorie content and price following an o...

The comparison of temporal methods

This work explores a new affective approach to projective mapping, based on consumers' choices or preferences. Two sessions, one week apart, were performed with the same consumers, using whole bread as a case study. Overall liking ratings (OL) were gathered in blind conditions and samples were also profiled by a trained panel using generic descript...

Projective mapping (PM), one of the most holistic product profiling methods in approach, is increasingly being used to uncover consumers' perception of products and packages. Assessors rely on a process of synthesis for evaluating product information, which would determine the relative importance of the perceived characteristics they use for mappin...

Ideas are essential for innovation and for the continuous renewal of a firm’s product offerings. Previous research has argued that online communities contain such ideas. Therefore, online communities such as forums, Facebook groups, blogs etc. are potential gold mines for innovative ideas that can be used for boosting the innovation performance of...

The aim of this study was to evaluate consumers' perception of a complex set of stimuli as aromatically enriched wines. For that, two consumer based profiling methods were compared, concurrently run with overall liking measurements: projective mapping based on choice or preference (PM-C), a newly proposed method, and check-all-that-apply (CATA) que...

Understanding the interaction of sensory and extrinsic product attributes in consumer preferences has been identified as one of the key pillars for raising the likelihood of food products’ success in the market. Over the course of the last decade there has been increased attention on research emphasizing a combination of these food-choice driving p...

Multi-way data arrays are becoming more common in several fields of science. For instance, analytical instruments can sometimes collect signals at different modes simultaneously, as e.g. fluorescence and LC/GC-MS. Higher order data can also arise from sensory science, were product scores can be reported as function of sample, judge and attribute. A...

The effect of curd cooking temperature (40 and 42°C), pressing, and ripening on textural and sensory properties of vacuum packaged semi-cooked ewe’s milk cheese was studied. Chemical and microbiological analyses, colour, texture, and sensory characteristics were determined. Curd cooking temperature and pressing influenced aw, moisture and salt-in-m...

We present the response-oriented sequential alternation (ROSA) method for multiblock data analysis. ROSA is a novel and transparent multiblock extension of the partial least squares regression (PLSR). According to a “winner takes all” approach, each component of the model is calculated from the block of predictors that most reduces the current resi...

In many areas of science multiple sets of data are collected pertaining to the same system. Examples are food products which are characterized by different sets of variables, bio-processes which are on-line sampled with different instruments, or biological systems of which different genomics measurements are obtained. Data fusion is concerned with...

In many areas of science multiple sets of data are collected pertaining to the same system. Examples are food products which are characterized by different sets of variables, bio-processes which are on-line sampled with different instruments, or biological systems of which different genomics measurements are obtained. Data fusion is concerned with...

Two different strategies for investigating individual differences among consumers in choice experiments using the Mixed Logit Model are compared. The study is based on a consumer study of iced coffees in Norway. Consumers (n = 102) performed a choice task of twenty different iced coffee profiles varying in coffee type, production origin, calorie co...

The focus of the present paper is to propose and discuss different procedures for performing variable selection in a multi-block regression context. In particular, the focus is on two multi-block regression methods: Multi-Block Partial Least Squares (MB-PLS) and Sequential and Orthogonalized Partial Least Squares (SO-PLS) regression. A small simula...

The authors compare two conjoint analysis approaches eliciting consumer preferences among different product profiles of iced coffees in Norway: Rating-based and Choice-based conjoint experiments. In the conjoint experiments, stimuli were presented in the form of mock-up pictures of iced coffees varying in coffee type, production origin, calorie con...

Acceptance of a product by a consumer may result from a convoluted interplay between product attributes and individual characteristics of that consumer. Different methods that systematically combine product properties with consumer groups segmented on such characteristics have provided unprecedented insight, but ignore heterogeneity in acceptance w...

The authors compare two conjoint analysis approaches eliciting consumer preferences among different product profiles of iced coffees in Norway: Rating-based and Choice-based conjoint experiments. In the conjoint experiments, stimuli were presented in the form of mock-up pictures of iced coffees varying in coffee type, production origin, calorie con...

The authors compare two conjoint analysis approaches eliciting consumer preferences among different product profiles of iced coffees in Norway: Rating-based and Choice-based conjoint experiments. In the conjoint experiments, stimuli were presented in the form of mock-up pictures of iced coffees varying in coffee type, production origin, calorie con...

The aim of this paper is to study the effects of exposure to different types of chocolate on consumer responses. Three different chocolates were considered, one established market leader and two new chocolates from a product development project. A group of consumers participated two times in a central location test and were exposed to different cho...

The aim of the present work is to extend the Sequentially Orthogonalized- Partial Least Squares (SO-PLS) regression method, usually used for continuous output, to situations where classification is the main purpose. For this reason SO-PLS discriminant analysis will be compared with other commonly used techniques such as Partial Least Squares-Discri...

To achieve product success in the market, it is important to understand the interplay between sensory and non-sensory product attributes, since both dimensions must be optimised during the product development process. Contextual factors have been shown to affect the outcome of acceptance studies, and it is important that consumer responses to food...

This paper discusses the use of the new path modelling approach based on Sequential Orthogonalised PLS regression within the context of consumer science. The method is based on splitting the estimation process into a sequence of modelling steps for each dependent block versus its predictive blocks. Focus will be on how the method can be used to com...

Data from descriptive sensory analysis are essentially three-way data with assessors, samples and attributes as the three ways in the data set. Because of this, there are several ways that the data can be analysed. The paper focuses on the analysis of sensory characteristics of products while taking into account the individual differences among ass...

The main objective of this paper is to describe and discuss the use of classification trees in consumer studies. Focus will be given to the use of the method in relating segments of consumers, based on their acceptance pattern, to additional consumer characteristics, including attitudes, habits and demographics variables. Advantages of the method i...

This paper presents a comparison of different methods for analyzing designed experiments. The methods used are based on PCA, PLS and ANOVA, used either separately or in combination. Special emphasis will be on how to obtain information about medium and less important factors in the presence of very dominating ones. It will be shown that this could...

Norvegia cheese samples (459 in total) were analysed during maturation using a broad range of analytical methods. From eight to 24 and 40 weeks there was a highly systematic development in chemical and sensory attributes. Chemical data and Fourier transform infra-red (FTIR) measurements gave almost equivalent validation results in prediction of sen...

In consumer studies the collected consumer data are often of different nature (demographic variables, attitudes and habits). Usually these data are considered all together when modelling consumer acceptance patterns, even though there may exist interesting relations between groups of consumer characteristics. The objective of this paper is thus to...

The main aim of this study was to investigate consumer preferences for extrinsic attributes of iced coffee, explore consumers’ coffee consumption habits, find new market opportunities and segment consumers based on similar products preferences. A sample of 101 consumers of iced coffee was recruited during 2012 in Norway. Twelve iced coffee products...