Detection and Classification of Hydrolyzed Dairy Additives Using Electronic Noses

Abstract

In the present work the classification of dairies additives in a rapid and simple way, is proposed. Measurements were made by means of an Electronic Nose developed in our laboratories, which we named 'Patagonia E-Nose.' This E-Nose is composed of SnO2 sensors located in a thermally stabilized chamber which improves the repeatability of the measurements. Samples of various hydrolyzed dairies were measured using air as a reference gas. Then, the integrals of the signals were analyzed using a combination of different multivariate chemometric methods such as linear discriminant analysis (LDA), principal components analysis (PCA), back-propagation neural network and different classifiers. Also, different algorithms were implemented and compared by calculating the number of correctly classified samples of each method. 99.4% of correct classifications were obtained by using cross-validation and selecting the most appropriate algorithms. The results indicate that the samples were correctly classified through the implementation of a simple and low cost measurement protocol.