Biomedical Text Classification Using Augmented Word Representation Based on Distributional and Relational Contexts

Table 15

LSTM classification performance using embeddings over the PubMed_20k_RCT dataset.

Embeddings	Accuracy
	,		,		,		,
	Training	Validation	Training	Validation	Training	Validation	Training	Validation
GloVe_C	0.7605	0.7589	0.7826	0.7522	0.7601	0.7583	0.7724	0.7719
GloVe_W	0.7589	0.7566	0.7708	0.7680	0.7612	0.7559	0.7726	0.7540
GloVe_Merged	0.7309	0.7295	0.7407	0.7378	0.7324	0.7201	0.7433	0.7405
CE	0.7667	0.7606	0.7785	0.7554	0.7687	0.7596	0.7856	0.7754
WE	0.7571	0.7559	0.7819	0.7731	0.7707	0.7587	0.7849	0.7699
Merged	0.7386	0.7297	0.7771	0.7497	0.7431	0.7397	0.7592	0.7486

Bold means the best performance in the case of each dataset.