Discrimination of Honeys by Electronic Tongue and Different Analytical Techniques

A study aimed at classification of honey samples from different floral origins by electronic tongue. The analytical data of 196 samples from different floral (eight) origins have been processed by Principal Component Analysis (PCA), Discriminant function analysis (DFA) and BP artificial neural network (BP-ANN). Using the electronic signal of sensors as input and floral origins as output, a twelve-hidden-layer BP Neural Network model was developed. It is shown that 100% of the training set and more than 98.4% of the testing set were classified by BP-ANN (network topology 7-12-12).