Table 6.

Evaluation using gene/protein name snippets from MEDLINE abstracts

	Dictionary			Lookup performance
Iter.	Ambiguity	Variability	Rule	Precision	Recall
0	5.797	12.479	(convert capital letters to lower case)	0.782	0.582
1	5.807	12.161	‘-’ → ‘’	0.766	0.603
2	5.811	12.025	‘ precursor’ → ‘’	0.767	0.611
3	5.812	11.941	‘,’ → ‘’	0.767	0.611
4	5.812	11.907	‘inc finger protein’ → ‘nf’	0.767	0.611
5	5.812	11.868	‘ isoform 1’ → ‘’	0.767	0.611
6	5.813	11.832	‘ isoform 2’ → ‘’	0.766	0.611
7	5.813	11.806	‘ isoform a’ → ‘’	0.766	0.611
8	5.813	11.781	‘ isoform b’ → ‘’	0.766	0.611
9	5.813	11.748	‘ containing protein’ → ‘containing’	0.766	0.611
10	5.813	11.730	‘ variant’ → ‘’	0.766	0.611
:	:	:	:	:	:
21	5.815	11.597	‘nterleukin’ → ‘l’	0.767	0.613
:	:	:	:	:	:
24	5.816	11.566	‘specific’ → ‘’	0.767	0.615
:	:	:	:	:	:
33	5.816	11.450	‘protein’ → ‘gene’	0.765	0.616
34	5.828	11.056	‘ gene’ → ‘’	0.765	0.619
:	:	:	:	:	:
38	5.829	11.016	‘ recepto’ → ‘’	0.767	0.623
:	:	:	:	:	:
44	5.830	10.970	‘ alph’ → ‘’	0.765	0.625
:	:	:	:	:	:
75	5.831	10.838	‘ i’ → ‘1’	0.766	0.626
:	:	:	:	:	:
84	5.831	10.790	‘ lpha’ → ‘’	0.766	0.627
:	:	:	:	:	:
86	5.831	10.782	‘ beta’ → ‘b’	0.767	0.630
:	:	:	:	:	:
100	5.832	10.732	‘ type’ → ‘’	0.767	0.633