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

Table 17

CNN-LSTM classification performance using embeddings over the PubMed_20k_RCT dataset.

Embeddings	Accuracy
	,		,		,		,
	Training	Validation	Training	Validation	Training	Validation	Training	Validation
GloVe_C	0.7275	0.7217	0.7564	0.7497	0.7246	0.7156	0.7509	0.7340
GloVe_W	0.7233	0.7126	0.7220	0.7278	0.7275	0.7175	0.7473	0.7379
GloVe_Merged	0.7107	0.7010	0.7121	0.7207	0.7190	0.7049	0.7327	0.7231
CE	0.7337	0.7186	0.7515	0.7440	0.7390	0.7195	0.7908	0.7553
WE	0.7303	0.7184	0.7532	0.7505	0.7418	0.7293	0.7960	0.7756
Merged	0.7273	0.7094	0.7507	0.7491	0.7347	0.7109	0.7785	0.7531

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