Title: Feature extraction of electronic nose for classification of indoor pollution gases based on kernel entropy component analysis

Authors: Jia Yan; Shukai Duan; Lidan Wang; Pengfei Jia; Tingwen Huang; Fengchun Tian; Kun Lu

Addresses: College of Electronic and Information Engineering, Southwest University, Chongqing 400715, China ' College of Electronic and Information Engineering, Southwest University, Chongqing 400715, China ' College of Electronic and Information Engineering, Southwest University, Chongqing 400715, China ' College of Electronic and Information Engineering, Southwest University, Chongqing 400715, China ' Department of Electrical and Computer Engineering, Texas A&M University at Qatar, P.O. Box 23874, Doha, Qatar ' College of Communication Engineering, Chongqing University, Chongqing 400044, China ' China International Engineering Consulting Corporation, No. 32 Che Gong Zhuang Xilu, Beijing 100048, China

Abstract: Feature extraction is important for electronic nose (E-nose), when it is used to classify different gases or odours. A novel feature extraction technique of E-nose based on kernel entropy component analysis (KECA) is presented in this paper. KECA is integrated with Renyi entropy and extracts the features from the kernel Hilbert space by projecting the input dataset onto the kernel principal component analysis (KPCA) axes that preserve the most Renyi entropy. Besides KECA, independent component analysis and KPCA are also used to deal with the original feature matrix of four different indoor pollution gases acquired by E-nose. Experimental results prove that the classification accuracy of KECA is better than other considered techniques.

Keywords: feature extraction; electronic nose; E-nose; Renyi entropy; KECA; kernel entropy component analysis; indoor pollution gas.

DOI: 10.1504/IJISTA.2017.084224

International Journal of Intelligent Systems Technologies and Applications, 2017 Vol.16 No.2, pp.140 - 152

Accepted: 14 Sep 2016
Published online: 21 May 2017 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article