FEDRR: fast, exhaustive detection of redundant hierarchical relations for quality improvement of large biomedical ontologies

Table 7 Summary of the results for randomly generated ontologies

(N,E,C_min,C_max)	# Layers	TC	RR	RR%	T(ms)
(500,000, 550,000, 2, 5)	12	5,957,690	6	0.0001	5,847
(500,000, 550,000, 2, 10)	8	4,036,792	12	0.0003	5,212
(500,000, 550,000, 2, 20)	6	3,000,751	7	0.0002	5,637
(500,000, 600,000, 2, 20)	6	3,567,691	13	0.00036	5,449
(500,000, 700,000, 2, 20)	6	4,190,494	34	0.00081	5,332
(500,000, 900,000, 2, 20)	6	7,104,749	109	0.00153	8,183
(500,000, 1,300,000, 2, 20)	6	25,934,499	1,404	0.0054	33,235
(1,000,000, 1,200,000, 2, 20)	7	7,071,813	21	0.0003	18,051
(1,000,000, 1,400,000, 2, 20)	7	8,694,703	133	0.0015	11,582

N: number of concepts, E: number of is-a relations, C_min/C_max: minimum/maximum number of children a node can have, TC: number of transitive closure pairs, RR: number of redundant is-a relations, T(ms): time taken in milliseconds

ISSN: 1756-0381