Performance evaluation of classifier techniques to discriminate odors with an E-Nose

The foremost objective of this research is to discriminate odors with an E-Nose using various classifier technique based on statistical machine learning theory. The goal of learning is prediction. The learning problem comprises of the purpose that is used to map between the input and the output in a predictive manner, such that the output can be predicted from future input using the function that is learnt. The TGS series gas sensors (TGS 2602, TGS2620, TGS2201A and TGS2201B) are employed which are stimuli to several indoor air contaminants. With the help of the classifier techniques we will differentiate the gases and other organic compounds. As to this four techniques have been used as a classifier in this context such as PCA, LDA, SVM and NBC. The learning techniques are categorized into supervised and non-supervised. The comparisons among various techniques have been drawn to differentiate the techniques on the basis of classification accuracy.