Common components and specific weight analysis and multiple co‐inertia analysis applied to the coupling of several measurement techniques

Abstract

The present paper compares two multiblock techniques: the Common Components and Specific Weights Analysis (CCSWA) and the Multiple Co-inertia Analysis (MCoA). Both methods are used to (1) to investigate the relationships among various data tables and (2) to extract latent variables from information of different nature, reflecting different facets of a food product. Our objective is to study the ability of these methods to extract, from a set of data tables, latent characteristics which are representative of the whole modifications brought to a complex system (food product) by a modification of a given process factor. The comparison of these methods is based on the investigation of their conceptual framework by particularly highlighting new properties of CCSWA. Moreover, the two techniques of analysis are compared on the basis of a case study in cheese processing where each cheese sample is described by different kinds of measurements.

Shedding new light on Hierarchical Principal Component Analysis

“…It was also successfully applied in several domains of application and in particular in chemometrics. For a more recent investigation of CCSWA, we refer to Mazerolles et al [15] and Hanafi et al [4,16].…”

“…We refer to a paper by Janné et al [19] for an application of HPCA within the framework of spectroscopic data. CCSWA was applied for the coupling of several instrumental measurements [4,15,23]. We herein use a simple case study pertaining to sensory profiling to illustrate the methods of analysis and we aim at focusing on the complementarities of HPCA and CCSWA.…”

“…Several methods have been proposed for the statistical treatment of multiblock datasets. For a comprehensive review, we refer to Hanafi et al [4] and Tenenhaus et al [5]. In this paper, we focus on Hierarchical Principal Component Analysis (HPCA), which was introduced by Wold et al [6].…”

Improved classification of fused data: Synergetic effect of partial least squares discriminant analysis (PLS-DA) and common components and specific weights analysis (CCSWA) combination as applied to tomato profiles (NMR, IR and IRMS)

“…In later studies CCSWA was successfully applied to different data sets of food products [20][21][22][23] as well as food packaging materials [24]. The results have shown that CCSWA extracts relevant information more efficiently than classical single-block chemometric approaches [20,24].…”

“…The results have shown that CCSWA extracts relevant information more efficiently than classical single-block chemometric approaches [20,24]. However, in all these studies CCSWA was used as an exploratory tool and not for prediction of class membership of new samples [20][21][22][23][24].…”