Table 6.

Parameter sensitivity analysis: model performance across different hyperparameter settings.

Parameter	Setting	Accuracy (%)	Recall (%)	F1 score (%)
GCN layer depth	2	89.34	88.12	88.72
	3	92.56	91.78	92.17
	4	91.01	89.95	90.47
Attention dimension	64	90.12	89.45	89.78
	128	93.01	92.43	92.72
	256	91.45	90.11	90.77
Domain regularization weight	0.1	90.45	89.67	90.05
	0.5	93.12	92.56	92.83
	1.0	91.23	90.34	90.78

The values in bold are the best values.