Classification and pattern recognition algorithms applied to E-Nose

Electronic noses (E-Nose) are seen to be good substitute to human or animal nose for food, and fruit quality identification. It is also used for explosive and chemical identification. We have generated typical E-Nose data to compare the existing algorithms in terms of training, and testing/validation. We have observed that k-nearest neighbor (k-NN) algorithm, support vector machine (SVM) machine learning algorithms; and radial basis function (RBF), and generalized regression neural networks (GRNNs) shows good odor detection performance. In terms of speed GRNN tops compared to other methods.