Table 3.

Exploring the impact of model components through ablation study on CMU-MOSEI and MIMIC-IV datasets.

Model	CMU-MOSEI dataset				MIMIC-IV dataset
	Accuracy ↑	Precision ↑	Recall ↑	F1 score ↑	Accuracy ↑	Precision ↑	Recall ↑	F1 score ↑
w./o. Attention-Guided Diagnosis	89.1 ± 0.3	87.4 ± 0.4	85.9 ± 0.3	86.5 ± 0.3	91.0 ± 0.3	88.3 ± 0.4	86.8 ± 0.3	87.3 ± 0.3
w./o. Domain-Adaptive Learning	88.5 ± 0.4	87.0 ± 0.3	85.4 ± 0.3	86.0 ± 0.3	90.6 ± 0.3	88.0 ± 0.3	86.5 ± 0.4	87.0 ± 0.3
w./o. Uncertainty-Aware Prediction	88.9 ± 0.3	87.2 ± 0.4	85.6 ± 0.3	86.2 ± 0.3	90.8 ± 0.3	88.1 ± 0.3	86.6 ± 0.4	87.1 ± 0.3
Ours	90.3 ±0.3	88.2 ±0.3	86.9 ±0.4	87.5 ±0.3	92.1 ±0.3	89.7 ±0.3	88.2 ±0.4	88.8 ±0.3

Bold values indicate the best performance in each column.