Table 2.

Overall performance of the tested models.

	LR	SVM	XGB	MLP
Demographics	0.668 [0.665–0.671]	0.659 [0.658–0.660]	0.675 [0.674–0.676]	0.655 [0.652–0.658]
TF-IDF	0.682 [0.678–0.686]	0.692 [0.691–0.693]	0.673 [0.670–0.676]	0.630 [0.620–0.640]
LDA-75	0.661 [0.660–0.662]	0.663 [0.662–0.664]	0.634 [0.633–0.635]	0.649 [0.646–0.652]
Demographics + TF-IDF	0.687 [0.679–0.695]	0.698 [0.697–0.699]	0.688 [0.687–0.689]	0.631 [0.621–0.641]
Demographics + LDA	0.685 [0.684–0.686]	0.685 [0.684–0.686]	0.679 [0.677–0.681]	0.643 [0.628–0.658]
Demographics + TF-IDF + LDA-75	0.696 [0.695–0.697]	0.701 [0.699–0.703]	0.688 [0.687–0/689]	0.626 [0.607–0.645]
Demographics + TF-IDF + LDA-75 + PrPA	0.726 [0.725–0.727]	0.721 [0.720–0.722]	0.706 [0.707–0.705]	0.628 [0.612–0.644]

The bold font corresponds to the best performing model for a given set of features (in the leftmost column).