Least Square Support Vector Machines in Combination with Principal Component Analysis for Electronic Nose Data Classification

In this paper, an electronic nose data classification approach based on least square support vector machines (LS-SVM) in combination with principal component analysis (PCA) is investigated. The electronic nose data are first converted into PCA, where the data are projected from a high dimensional space into a low dimensional space, preferably two or three dimensions. Then the resulting features from the PCA are sent into the LS-SVM classifier in order to recognize the gas category. The performance of the proposed approach is validated by cross-validation technique. An experiment has been demonstrated by using coffee data from different types of coffee blends. Experimental results show that the LS-SVM in combination with PCA is an effective technique for the classification of electronic nose data.