Влияние фильтрации на классификацию спектральных данных

Abstract

In the present study, we considered the influence of filtration on classification of model absorption spectra of exhaled air of healthy volunteers and patients with obstructive chronic lung disease with additive noise. The classification was carried out by the Support Vector Machine (SVM). According to the analysis the median filter negatively affects on data classification, the Savitzky-Golay filter increases specificity, however reduces sensitivity, the Second Order Butterworth Filter increases specificity and sensitivity.