Table 4.

Top 20 most frequent keywords (sorted in ascending order of Year and descending order of Count)

Context	Phase	Keyword	Year	Count	Density	Between
Core	P1	unified medical language system	1991	63	2.172	0.120
		Medline	1993	19	0.704	0.010
		System	1994	87	3.346	0.140
		Information	1994	62	2.385	0.240
		Vocabulary	1994	22	0.846	0.050
		Umls	1995	160	6.400	0.280
	P2	Terminology	1998	62	2.818	0.170
		Knowledge	1998	42	1.909	0.050
		Database	1998	35	1.591	0.040
		information retrieval	1998	29	1.318	0.060
		medical language system	1999	49	2.333	0.060
		Representation	1999	22	1.048	0.020
		natural language processing	2003	86	5.059	0.200
		Ontology	2003	72	4.235	0.110
		Text	2004	61	3.813	0.150
		information extraction	2004	22	1.375	0.090
		Network	2004	18	1.125	0.050
		Classification	2008	23	1.917	0.090
	P3	Extraction	2011	21	2.333	0.060
		machine learning	2011	20	2.222	0.100
Expanded	P1	System	1992	1196	42.714	0.040
		Information	1992	801	28.607	0.060
		Database	1994	667	25.654	0.060
		information retrieval	1994	362	13.923	0.030
		Terminology	1994	348	13.385	0.040
		Ontology	1995	1084	43.360	0.080
		Knowledge	1995	556	22.240	0.050
		natural language processing	1995	550	22.000	0.020
		Classification	1995	409	16.360	0.080
		Care	1995	345	13.800	0.050
		Umls	1995	327	13.080	0.040
		Disease	1996	293	12.208	0.040
		Model	1997	443	19.261	0.050
	P2	Text	2001	489	25.737	0.020
		Tool	2001	329	17.316	0.020
		Extraction	2002	292	16.222	0.020
		text mining	2003	312	18.353	0.010
		Gene	2003	296	17.412	0.010
		Identification	2004	291	18.188	0.020
		electronic health record	2006	350	25.000	0.040