HESML: a real-time semantic measures library for the biomedical domain with a reproducible survey

BMC Bioinformatics

Table 12 Experimental confirmation of the \(k{\bar{E}}^2_{C}\) factor impacting the linear scalability of AncSPL for non-tree-like ontologies (\(TC_3\)) shown in Fig. 3

Ontology	\({\bar{E}}_{C}\)	\(\widehat{k{\bar{E}}^2_{C}}\) \(({\mu }s)\)	\({\bar{E}}^2_{C}/{\bar{E}}^2_{WN}\)	\(\widehat{{\bar{E}}^2_{C}/{\bar{E}}^2_{WN}}\)
SNOMED-CT	72.02	1.191	7.79	5.39
GO	31.14	0.3277	1.46	1.48
WordNet (WN)	25.80	0.2210	1	1

First column shows the average number of adjacent nodes per ancestor set for each node in ontology C, denoted by \({\bar{E}}_{C}\). Second column shows the estimated value for the factor \(k{\bar{E}}^2_{C}\) in \(TC_3\) obtained by fitting the scalability plot shown in Fig. 3 to the line \(t_{{\mu }s} = \alpha + (k{\bar{E}}^2_{C})N\). Then, third and fourth columns compare the theoretical and experimental expected ratios between the time complexity (slope) of two different ontologies using WordNet (WN) as baseline

ISSN: 1471-2105