Two-step PLS regression for L-structured data: an application in the cosmetic industry

Vinzi, Vincenzo Esposito; Guinot, Christiane; Squillacciotti, Silvia

doi:10.1007/s10260-006-0028-2

Cited by 15 publications

(5 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The L‐PLS regression approach introduced by Martens et al (2005) is based on one single analysis combining all the three blocks of data together (i.e., sensory properties, consumers' degree of liking ratings, and consumers' attributes; Vinzi et al, 2007). The matrices

X

and

Z

are centered for properties and attributes respectively, while matrix

Y

is supposed to be centered with respect to both its rows and its columns (double centered).…”

Section: Theory: Statistical Methodsmentioning

confidence: 99%

“…L-shape data can be analyzed in different ways, see for example, Smilde et al (2022) and Vinzi et al (2007). Here, we will focus on a one-step approach called L-Partial Least Squares (L-PLS) regression, and a two-step procedure (TSP) using standard Partial Least Squares (PLS) regression methods along the horizontal and vertical direction in the L-shape separately.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Comparison of different ways of handling consumer segments using L‐shape data

Nguyen

Asioli

Varela

et al. 2023

Journal of Sensory Studies

View full text Add to dashboard Cite

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 attributes. Data were analyzed using two different approaches. In approach one, the one‐step simultaneous L‐Partial Least Square (L‐PLS) Regression with average consumer liking to represent the segments was used, while approach two was based on a two‐step procedure (TSP) based on Partial Least Square (PLS) Regression using dummy variables to represent the segments. The methods were compared in terms of interpretations, flexibility, and outcomes. Methodological implications, recommendations, and future research avenues are discussed. Practical Applications This manuscript has been devoted to two different ways of handling segmentation in L‐shape data of consumer liking, sensory properties, and consumer attributes. Overall, both L‐PLS and TSP approaches provide similar interpretation of results. The TSP approach, however, has the advantage of interpretating the horizontal and vertical direction in the L separately using standard regression methods. It is of interest of product development and marketing activities to identify which food product characteristics are important for consumer preferences and to better understand the characteristics of the consumers (e.g., socio‐demographics) that drive the consumer acceptance of the different products.

show abstract

X

and

Z

are centered for properties and attributes respectively, while matrix

Y

is supposed to be centered with respect to both its rows and its columns (double centered).…”

Section: Theory: Statistical Methodsmentioning

confidence: 99%

mentioning

confidence: 99%

Comparison of different ways of handling consumer segments using L‐shape data

Nguyen

Asioli

Varela

et al. 2023

Journal of Sensory Studies

View full text Add to dashboard Cite

show abstract

“…In recent years, a number of data analysis approaches have been suggested to handle L-shaped 101 data set (see e.g. Vinzi, Guinot, & Squillacciotti, 2007). The first part of the present sub-section 102 will be devoted to the two-step approach (PLS regression, see e.g.…”

Section: L-shaped Data 100mentioning

confidence: 99%

Comparison of different ways of handling L-shaped data for integrating sensory and consumer information

Asioli

Nguyen

Varela

et al. 2022

Food Quality and Preference

View full text Add to dashboard Cite

Different approaches for handling L-shaped data are compared for the first time in a study 17 conducted with Norwegian consumers. Consumers (n = 101) valuated eight different yoghurt 18 profiles varying in three intrinsic attributes such as viscosity, particle size, and flavour intensity 19 following a full factorial design. Sensory attributes, consumers' liking ratings, and consumer 20 attributes were collected. Data were analysed using two different approaches of handling L-21 shaped data: approach one used two-step Partial Least Square (PLS) Regression using L-shaped 22 data including the three blocks such as sensory attributes, consumers' liking ratings, and 23 consumer attributes, while approach two was based on one-step simultaneous L-Partial Least 24 Square (L-PLS) Regression model of the same three blocks of data. The different approaches 25 are compared in terms of centering, step procedures, interpretations, flexibility, and outcomes. 26 Methodological implications and recommendations for academia and future research avenues 27 are outlined. 28

show abstract

“…According to the lines introduced by Wold et al (1983), several researchers have been taken all these matrices into account by means of Partial Least Squares (PLS) regression methods (Martens et al, 2005;Esposito Vinzi et al, 2007. This approach has been specifically applied to sensory data, for clustering and classification purposes (Vigneau and Qannari, 2003), as L-PLS regression by Endrizzi (2008) and Plaehn and Lundahl (2006), among others.…”

Section: Structure Of Data and Model Specificationmentioning

confidence: 99%

A multi-product approach for detecting subjects’ and objects’ covariates in consumer preferences

Capecchi

Endrizzi

Gasperi

et al. 2016

British Food Journal

View full text Add to dashboard Cite

Purpose – A different framework based on a parametric version of the process generating the hedonic scores is adopted. More precisely, a probability distribution for ordinal responses is proposed as a mixture of two components, denoted as feeling (as expressed preference) and uncertainty component (as inherent indecision). The purpose of this paper is to analyse the effect of covariates on the consumers’ behaviour pattern according to a statistical model. Design/methodology/approach – Sample data come from a multidisciplinary research aimed to improve the quality and marketability of soft fruits. Then, a stochastic model with subjects’ and objects’ covariates is built and the interpretation of significant results is discussed. Findings – The joint effects of personal characteristics and chemical contents of juice on the hedonic scores given by consumers are examined and graphically depicted by means of a significant model. Originality/value – The paper suggests a multi-product approach to expressed hedonic scores by means of a generalization of CUB models

show abstract

Two-step PLS regression for L-structured data: an application in the cosmetic industry

Abstract: Partial least squares (PLS) regression, Preference data, External information, L-structures,

Cited by 15 publications

References 9 publications

Comparison of different ways of handling consumer segments using L‐shape data

Comparison of different ways of handling consumer segments using L‐shape data

Comparison of different ways of handling L-shaped data for integrating sensory and consumer information

A multi-product approach for detecting subjects’ and objects’ covariates in consumer preferences

Contact Info

Product

Resources

About