An Ontology-Based Semantic Similarity Measure Considering Multi-Inheritance in Biomedicine

Table 3

Correlations across measures for physicians, coders, and both on 29 pairs of Dataset 1 and correlations across measures for human on 34 pairs of Dataset 2 using SNOMED CT as ontology.

Measure	Dataset 1			Dataset 2
	Physicians	Coders	Both	Human
Path length	0.36	0.51	0.48	0.586
Leacock and Chodorow	0.35	0.50	0.47	0.677
Wu and Palmer	N/A	0.29	N/A	0.686
Li et al.	N/A	0.37	N/A	0.694
Choi and Kim	N/A	0.15	N/A	0.440
Al-Mubaid and Nguyen (SemDistA&N)	N/A	0.66	N/A	0.735
Montserrat Batet et al. (SimB&S)	0.60	0.79	0.73	N/A
Resnik	0.45	0.62	0.55	N/A
Lin	0.60	0.75	0.69	N/A
Jiang and Conrath	0.45	0.62	0.55	N/A
Context vector (1 m notes, diagnostic section)	0.84	0.75	0.76	N/A
Context vector (1 m notes, all sections)	0.62	0.68	0.69	N/A
Context vector (100,000 notes, diagnostic section)	0.56	0.59	0.60	N/A
Context vector (100,000 notes, all sections)	0.41	0.53	0.51	N/A
Our measure	0.67	0.77	0.75	0.774