Figure 2
From: Reduced levels of pulmonary surfactant in COVID-19 ARDS
A principal component analysis (PCA) score plot with projection of the data onto the span of the principal components (PC). The spectral matrices were decomposed into two principal components (PCs) by applying singular value decomposition, resulting in a score plot of the first and second PC. Each PC is a linear combination of the wavenumbers of spectra. Before applying PCA, the spectra were baseline corrected (Whittaker smoother) and subsequently normalized by applying standard normal variate methods to avoid intensity variations in the spectra due to the deposition. Two parameters, lambda (λ = 105) and penalty (P = 10–3) are involved related to the smoothness of the fit and the penalty imposed to the points giving positive residuals in the fit.
