Hybrid electronic tongue based on multisensor data fusion for discrimination of beers

This paper reports the use of a hybrid electronic tongue based on data fusion of two different sensor families, applied in the recognition of beer types. Six modified graphite-epoxy voltammetric sensors plus 15 potentiometric sensors formed the sensor array. The different samples were analyzed using cyclic voltammetry and direct potentiometry without any sample pretreatment in both cases. The sensor array coupled with feature extraction and pattern recognition methods, namely Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA), was trained to classify the data clusters related to different beer varieties. PCA was used to visualize the different categories of taste profiles, while LDA with leave-one-out cross-validation approach permitted the qualitative classification. The aim of this work is to improve performance of existing electronic tongue systems by exploiting the new approach of data fusion of different sensor types.