Figure 5
From: HHT diagnosis by Mid-infrared spectroscopy and artificial neural network analysis
Intra class principle component analysis (PCA) for the HHT-negative and HHT-positive group (class). The four best eigenvectors for the HHT-negative (blue circles) and the HHT-positive (red crosses) group were seleted separately for region 1, 3150–2750 cm-1, and region 2, 1850–750 cm-1. Projecting the data onto the eight eigenvectors, resulting in 2x8 features (PC1-PC4 of the positive and negative group and of region 1 and 2), shows that the HHT-positive and HHT-negative groups are multimodal (multiple clusters) with an overlapping mode between the two classes.