Table 6.

Overall and domain-specific Precision, Recall, and F1 scores for our models

	Precision	Recall	F1
Aggregate Cosine Similarity Scores	0.626	0.692	0.657
MLP Baseline (No MWEs)	0.816	0.830	0.823
RBF Baseline (No MWEs)	0.795	0.808	0.801
MLP (w/ MWEs)	0.821	0.835	0.828
Appearance	0.953	0.825	0.884
Interpersonal	0.843	0.897	0.869
Mood	0.723	0.816	0.767
Occupation	0.945	0.834	0.886
Substance	0.898	0.946	0.921
Thought Content	0.830	0.685	0.751
Thought Process	0.792	0.878	0.833
Other	0.509	0.614	0.557
RBF (w/ MWEs)	0.814	0.799	0.806
Appearance	0.952	0.803	0.871
Interpersonal	0.929	0.882	0.905
Mood	0.748	0.759	0.754
Occupation	0.956	0.847	0.898
Substance	0.826	0.927	0.874
Thought Content	0.866	0.685	0.765
Thought Process	0.958	0.818	0.883
Other	0.405	0.411	0.408