A Novel Pre-Processing Technique for Original Feature Matrix of Electronic Nose Based on Supervised Locality Preserving Projections.

Pengfei Jia,Li Wang,Tailai Huang,Jia Yan,Shukai Duan,Lidan Wang

doi:10.3390/s16071019

Pengfei Jia, Li Wang + Show 4 more

Open Access

https://doi.org/10.3390/s16071019

Copy DOI

Journal: Sensors (Basel, Switzerland)	Publication Date: Jun 30, 2016
Citations: 41	License type: CC BY 4.0

Affiliation: Southwest University

Abstract

An electronic nose (E-nose) consisting of 14 metal oxide gas sensors and one electronic chemical gas sensor has been constructed to identify four different classes of wound infection. However, the classification results of the E-nose are not ideal if the original feature matrix containing the maximum steady-state response value of sensors is processed by the classifier directly, so a novel pre-processing technique based on supervised locality preserving projections (SLPP) is proposed in this paper to process the original feature matrix before it is put into the classifier to improve the performance of the E-nose. SLPP is good at finding and keeping the nonlinear structure of data; furthermore, it can provide an explicit mapping expression which is unreachable by the traditional manifold learning methods. Additionally, some effective optimization methods are found by us to optimize the parameters of SLPP and the classifier. Experimental results prove that the classification accuracy of support vector machine (SVM combined with the data pre-processed by SLPP outperforms other considered methods. All results make it clear that SLPP has a better performance in processing the original feature matrix of the E-nose.

Highlights

An electronic nose (E-nose) is an expert system which is composed of an array of gas sensors as well as a corresponding artificial intelligence algorithm
To verify the effectiveness of supervised locality preserving projections (SLPP), the original feature matrix of wound infection is processed by principal component analysis (PCA) [28], Fisher discriminant analysis (FDA) [29] and kernel FDA
Further experimental results prove that the classification accuracy of the E-nose increases when SLPP is used to pre-process the original feature matrix of wound infection data, and it reduces the dimension of points from 15 to 7, which can greatly decrease the computational complexity of classifier

Summary

Introduction

An electronic nose (E-nose) is an expert system which is composed of an array of gas sensors as well as a corresponding artificial intelligence algorithm. The sensor array of the E-nose has a characteristic of cross-sensitivity, namely different units in the array will make responses when facing the same smell, which can effectively avoid the decision-making risk brought by the single sensor. The original feature matrix which is extracted from the response of sensors is redundant, and much key information which plays a crucial role in helping the E-nose make the right judgment is submerged. Previous work has confirmed that the E-nose can be used to detect wounds of patients in labs and it is feasible for the E-nose to detect bacteria including the in the investigation of bacterial volatile organic compounds (VOCs) from cultures and from swabs taken from wound-infected patients [8,9,10,11]. There are many kinds of pathogenic bacteria which can lead to wound infection. The same sampling experiment on one pathogenic bacterium must be repeated many times to make the

Results

Discussion

Conclusion