Skip to main content

Table 4 Exemplifying results for Experiment I: Top twelve ranked candidate terms (highest cosine similarity) from the word embeddings created with CBOW and Skip-gram for the target term “KLF7” that appears in the abstract of the PubMed article with ID = 23,468,932

From: Deep learning meets ontologies: experiments to anchor the cardiovascular disease ontology in the biomedical literature

 

CBOW

Skip-gram

Rank

Candidate terms from word embeddings

Cosine

Candidate terms from word embeddings

Cosine

1

MoKA

0.376371

Prrx2

0.601920

2

pluripotency-associated_genes

0.335113

Klf7

0.592946

3

Sp1_regulates

0.334092

Klf7(−/−)

0.590523

4

LOC101928923

0.333423

RXRG

0.589875

5

p107_dephosphorylation

0.331689

LOC101928923

0.585979

6

PU_1

0.329925

SOX-17

0.585295

7

histone_demethylase

0.323529

rs820336

0.585094

8

gene_promoter

0.321640

GLI-binding_site

0.581073

9

homeobox_protein

0.319997

Tead2

0.580012

10

histone_arginine

0.315875

hHEX

0.579868

11

transfated

0.314202

ACY-957

0.579542

12

are_unable_to_repress

0.313112

ETS1

0.577272