A fuzzy entropy based neural network classifier for odor identification of alcoholic beverages using tin oxide sensor array

This paper presents a novel method to odor based identification of alcoholic beverages using steady-state responses of a thick film tin oxide sensor array exposed to four different types of whiskies. A neural classifier designed to perform the identification task was trained by incorporating the class information in the training data set in the form of fuzzy entropies of the respective classes. The performance of the proposed classifier has been compared with that of those reported earlier, which generally employed fuzzy membership values to generate class information. The use of fuzzy entropy measure resulted in better identification of the alcoholic beverages as compared to those which are based on fuzzy membership representation. Fuzzy entropy representation also resulted in precise identification of the alcoholic beverages by using reduced number of sensors in the array.