Medical Text Classification Using Hybrid Deep Learning Models with Multihead Attention

Table 11

Analysis of results with different dropout rates for the proposed hybrid BiGRU model.

Dropout rate	Dataset	Accuracy (%)	Precision (%)	Recall (%)	F1-score (%)
0.2	Hallmarks	61.88	57.19	58.84	58.00
	AIM	89.72	83.58	82.97	83.27
0.3	Hallmarks	71.51	66.73	64.70	65.69
	AIM	62.39	57.18	55.23	56.18
0.4	Hallmarks	72.82	71.56	70.92	71.23
	AIM	94.22	89.87	88.21	86.12
0.5	Hallmarks	72.45	68.97	67.88	68.42
	AIM	81.61	75.98	74.92	75.44