Skip to main content

Table 5 Collection of pre-trained word embedding (WE and WEC) models and ontology-based vector models (OVM) evaluated in a previous series of experiments [58,59,60] by using the Java classes implementing their evaluation

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

WN

Family

Word embedding model

Yes

WEC

Attract-repel [127]

No

WE

FastText [128]

No

WE

GloVe [129]

No

WE

CBOW [130]

Yes

WEC

SymPatterns (SP-500d) [131]

No

WEC

Paragram-ws [132]

No

WEC

Paragram-sl [132]

Yes

WEC

Counter-fitting (CF) [133]

Yes

OVM

WN-RandomWalks [134]

Yes

OVM

WN-UKB [125]

Yes

OVM

Nasari [126]

  1. First column details which methods use WordNet during their training