Classification of beverages using a reduced sensor array

Abstract

An electronic tongue based on the sensor array of ion-selective electrodes (ISEs) combined with pattern recognition tools is applied to qualitative analysis of mineral waters and apple juices. The procedure of reducing of total number of the sensors in the array is described. Before and after reducing of number of the sensors, this device is capable of reliable discrimination between different brands of mineral waters and apple juices.Validation of this method is also confirmed using artificial neural networks as a classifier. The ability to recognize different brands of waters and juices after reduction of number of the sensors in the array is comparable to the ability to recognize the same samples by the system before reduction of number of the sensors.This method enables to specify quantitatively the capability of a sensor to discriminate between different classes of samples and can be used for any electronic tongues and noses applications.