Table 4.

Medication information predictions metrics results by models.

Label	RBS				BiLSTM + ELMo				BiLSTM + ELMo + RBS
	F-measure	Precision	Recall	F-measure		Precision	Recall	F-measure		Precision	Recall
Medication name	90.31	96.46	84.89	92.2		93.79	90.67	95.33		95.33	95.33
Medication class	13.33	87.5	7.22	62.3		66.28	58.76	64.36		61.9	67.01
Dosage	90.43	96.62	84.98	92.17		91.13	93.23	95.29		95.52	95.05
Frequency	86.13	98.89	76.28	92.8		93.3	92.31	92.24		93.04	91.45
Duration	48.89	49.25	48.53	82.17		86.89	77.94	78.79		81.25	76.47
Route	47.92	85.19	33.33	75.52		72.97	78.26	72.86		71.83	73.91
Condition	33.64	100	20.22	55.9		62.5	50.56	62.16		77.97	51.69

^aRBS: rule-based system

^bBiLSTM: bidirectional long short term memory.

^cELMo: embedding for language models.