An efficient chemometric strategy based on wavelet transform for preprocessing of multi-capillary column – electronic nose data

چکیده مقاله

The multi-capillary column (MCC) that is hyphenated with electronic nose (E-nose) is a novel sensitive and strong technique with ability to separate components of odors to facilitate detection ofthem and it can be used in many applications. Small signal with high noise levels in MCC-Enose dataset demonstrates a necessity of noise removal technique before chemometric analyses. Signal averaging is the classical and the simplest smoothing method that traditionally have been used for noise reduction. Recently, many new algorithms based on transformation for smoothing, denoising and compression of analytical signals have been proposed. The use of the wavelet denosing technique improves the signal-to-noise ratio (SNR) along with amplifying the instrument resolution. The wavelet denoising method was applied to amplify significant information and suppress noise level and remove spike in the raw MCCEnose dataset that represents a preference in comparing with another denoising method. The proposed methods have been performed on the real noisy raw MCC-Enose dataset with known peak positions as a typical noisy signal. The experimental results show that this method is robust denoising method and it also performs well in preserving signal features in MCC-Enose dataset