A Novel Multi-Dimensional Feature Extraction Framework of Data Sampled by Electronic Nose

Abstract

A feature extraction framework of electronic nose is proposed in this paper. Proposed framework is mainly composed of two parts; the one is the noise reduction subsystem which is described as the cascade connection between wavelet filter and moving average filter, the other is the feature extraction system which can be used to extract the integral feature and difference feature related to output of the noise reduction subsystem. In addition, details related to proposed framework including wavelet basis and its order and slide windows of moving average filter are deeply discussed too. Comparative experiment on real dataset is employed to demonstrate the effectiveness of proposed framework.