Research Article

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

Table 3

Word similarity performance with and .

Word embeddingsBioSimLexBioSimVerbMENMTurkRG65RWSCWSSimLex999TR9856UMNSRS-relUMNSRS-simVERB143WS353WS353RWS353S

GloVe_W0.395950.021020.086330.18420−0.108680.191590.331330.092550.148760.028630.119770.205320.271380.18560.28586
GloVe_C0.389420.024630.086130.18368−0.115690.192410.333560.090080.149760.029110.117950.204240.270770.188930.29353
GloVe_Merged0.392520.01670.089780.18971−0.11440.194160.328670.085790.148630.036660.132040.164670.265510.182940.29706
WE0.459360.139310.229850.341540.164160.175210.372850.170740.162250.216590.314500.235200.234070.185450.31335
CE0.459110.144520.227560.341590.154230.164080.379650.169410.168670.230400.315350.240510.266120.229270.33163
Merged0.468320.150000.231510.330070.123020.153210.373040.158820.166050.225270.308950.2082150.227190.193140.29428

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