LAUSR

An electronic tongue (e-tongue) based on an array of lipid/polymer membranes has been successfully developed for measuring the taste evolution of natural milk. The e-tongue consisted of 16 different lipid/polymer membranes combined with or without a pH sensor. The natural milk of bovine and goat were purchased from the local farming store in Malang-Indonesia. The taste measurement was carried out, from fresh (0 h) to stale (12 h), every two hours under room ambient without any treatment. The responses of the e-tongue were evaluated using a Principal Component Analysis (PCA) and a Linear Discriminant Analysis (LDA). From PCA results, the taste of both milk samples tends to change by time although some groups show a partial overlapping. LDA results show the high precision of the e-tongue in clustering taste evolution. The correctly classified groups after the cross-validation procedure were achieved 95.7 and 87.1% for bovine and goat milk, respectively. The improvement of the classification using LDA was obtained by adding data from a pH sensor of each measurement as 100 and 98.6% for bovine and goat milk, respectively. This work indicates that the lab-made e-tongue may be useful to predict the quality of natural milk for the food industry.