Information Quality Ratio as a novel metric for mother wavelet selection

Abstract

This study proposes Information Quality Ratio (IQR) as a new metric for mother wavelet selection in real-world applications. In mother wavelet selection, common metrics such as MSE and correlation coefficient highlight the morphological similarity as well as SNR focuses on enlarging signal power against noise power. Instead, IQR emphasizes that the reconstructed signal has to keep essential information from the original signal. Regarding mother wavelet selection problem, we also demonstrate the effect of wavelet transform at various decomposition levels to make a clear foundation of wavelet decomposition. In this study, IQR was used to determine the best-suited mother wavelet for electronic nose signals in beef quality classification. The experimental results show that IQR based mother wavelets have better capability to keep essential information from original signals than SNR, MSE, and correlation coefficient based mother wavelets. Moreover, it has better sensitivity to quantify the changes of signal structure than MSE and correlation coefficient.