Figure 10
Top: average \(F_{1}\) score over all study participants for classifiers trained on tail (blue) and head (red) data for increasing values of the threshold τ. The performance of a classifier trained on all data (green) and of a random baseline (cyan, dashed) are also reported, for comparison. Bottom: \(F_{1}\) score of each individual for thresholds \(\tau = 0.00001\) (left) and \(\tau = 0.00007\) (right). The x-axis represents the 156 participants. Individuals are sorted by increasing \(F_{1}\) score of the tail classifier