Comparison of Matlab’s multivariate analysis for qualitative and quantitative electronic taste sensing system

Abstract

A taste sensing system has been put in place by a team of Unimap’s researchers using an array of sensors and ‘off the shelf’ components. Qualitative and quantitative ‘tasting’ data (multi-dimensions) were subjected to multivariate analysis methods available in mathematical software Matlab (ver. 8.0). Comparative studies were done on the analysis of principal component analysis (PCA), common factor analysis (CFA), multi-dimensional scaling (MDS) and linear discriminant analysis (LDA). This study shows that some analysis methods are capable to discriminate basic taste qualities and differentiate different level of taste intensity.