Medical Text Classification Using Hybrid Deep Learning Models with Multihead Attention

Table 12

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

Optimizer	Dataset	Accuracy (%)	Precision (%)	Recall (%)	F1-score (%)
Adam	Hallmarks	70.52	74.78	73.16	73.96
	AIM	93.98	88.63	89.08	88.85
SGD	Hallmarks	69.72	65.23	61.83	63.48
	AIM	69.31	66.57	68.27	67.40
Nadam	Hallmarks	70.13	68.84	68.64	68.73
	AIM	80.62	75.96	72.56	74.22
AdaGrad	Hallmarks	65.15	62.46	55.81	58.94
	AIM	86.15	81.59	81.28	81.43