On the application of simple matrix methods for electronic tongue data processing: Case study with black tea samples

By Yaroshenko, Irina; Kirsanov, Dmitry; Kartsova, Liudmila; Bhattacharyya, Nabarun; Sarkar, Subrata & Legin, Andrey
Published in Sensors and Actuators B: Chemical NULL 2014

Abstract

The paper is devoted to the application of several simple mathematical methods that can be used for analysis of multisensor system data. RV and modified RV' matrix correlation coefficients, Tucker's congruency coefficients and canonical correlation analysis can be effectively employed to reveal the similarity degree shared between two different data sets obtained for the same samples. This approach can be particularly useful when the number of available samples is too small to construct and validate quantitative regression models, but still the researcher is interested in getting some numerical estimate of the data quality to judge on electronic tongue applicability for particular analytical tasks. As a case study for this work we analyzed the data from potentiometric and voltammetric electronic tongues acquired during the analysis of black tea samples. These samples were also assessed with standard physicochemical methods and professional sensory panel. The relationships between these four data sets were studied with various matrix correlation techniques.

Read Article » Back