SO-PLS as an exploratory tool for path modelling

The paper “SO-PLS as an exploratory tool for path modelling” written by Elena Menichelli, Trygve Almøy, Oliver Tomic, Nina Veflen Olsen and Tormod Næs has been published in Food Quality and Preference,

Abstract

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 combine individual variables or specific groups of variables in more general blocks with a broader interpretation, such as for instance consumer habits, attitudes and demographic variables. It will be discussed how the methods can solve some of the challenges met when starting a path modelling process. It will be explored how the method handles multidimensionality of the blocks and thus how the analysis is simplified, at least for explorative purposes, as compared to other more traditional path modelling approaches. The study shows that important relations are revealed in presence of different types of information like product attributes, consumer characteristics and acceptance.

Keywords:

Path modelling, SO-PLS, Orthogonalisation, Regression, Multi-block