Skip to main content

Table 3 Comparison of correlations across measures and reference standards with MeSH

From: Semantic similarity in the biomedical domain: an evaluation across knowledge sources

Benchmark

Concept Graph

Knowledge Based

 
  

PPR

Path finding

Intrinsic IC

Corpus IC Lin

  

Taxonomy

All

Wu & Palmer

Path, LCH

Path

LCH

Lin

 

Pedersen Coders N=29

msh

0.36

 

0.42

0.42

0.51

0.54

0.51

0.51

 

msh-umls

0.20

0.20

0.37

0.38

0.44

0.53

0.45

0.43

Pedersen Physicians N=29

msh

0.30

 

0.40

0.40

0.41

0.39

0.41

0.41

 

msh-umls

0.15

0.15

0.40

0.41

0.42

0.41

0.42

0.38

Pedersen Combined N=29

msh

0.31

 

0.42

0.42

0.46

0.43

0.46

0.46

 

msh-umls

0.16

0.16

0.41

0.42

0.44

0.45

0.45

0.41

Mayo N=61

msh

0.27

 

0.37

0.35

0.47

0.37

0.46

0.45

 

msh-umls

0.05

0.05

0.20

0.26

0.25

0.25

0.25

0.20

UMN similarity N=429

msh

0.27

 

0.25

0.26

0.32

0.29

0.33

0.33

 

msh-umls

0.18

0.18

0.26

0.25

0.29

0.29

0.30

0.29

UMN relatedness N=432

msh

0.35

 

0.33

0.33

0.41

0.36

0.41

0.42

 

msh-umls

0.18

0.18

0.35

0.34

0.34

0.35

0.34

0.34

  1. Abbreviations: LCH - Leacock & Chodorow, PPR – Personalized PageRank. Refer to Table 1 for concept graph descriptions.