Figure 3.

This figure demonstrates the 10 best features extracted from ReliefF algorithm for each dimension of personality as evaluated by the fusion scenario. Bivariate features are demonstrated by lines, whereas univariate features (here only betweenness centrality (BC)) are denoted by colored nodes. The color of each edge represents a specific brainwave (black is full spectrum (FS) from 4 to 45 Hz), while the maximum classification accuracy for each trait is written beneath the head plots. Finally, the solid lines denote features extracted from HVHA clips, while dashed lines denote features extracted by LVHA clips.