Figure 4

Evaluation of model performance. (a) Correlation of users to the trained model (green), other users (red) and the random classifier (magenta). The gray bar goes from 1 to the correlation of the user with themself (see Fig.Β 1d). (b) Mean correlations of models to all other models (blue), all users (red) along with the correlations of this model with the random classifier (yellow) and for comparison the userβs correlation from panel A (magenta).