Research Article

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

Table 5

Word similarity performance with and .

Word embeddingssBioSimLexBioSimVerbMENMTurkRG65RWSCWSSimLex999TR9856UMNSRS-relUMNSRS-simVERB143WS353WS353RWS353S

GloVe_W0.432070.026960.109920.25905−0.080430.197560.338680.105240.146930.072690.147840.257670.239130.153150.27493
GloVe_C0.433980.021420.111910.26266−0.074750.198110.34460.103840.147850.080830.155360.256380.239760.151990.27944
GloVe_M erged0.441340.022010.112450.25225−0.078520.198500.341790.099610.147270.083680.167400.230640.237760.154460.28936
WE0.478400.132750.206730.258420.184570.177880.384990.166910.152090.220260.326470.290770.340630.283520.40837
CE0.463820.130860.209280.261340.165820.211320.387830.167500.156910.246320.329260.293490.333430.244600.40477
Merged0.467630.173100.238660.267650.215090.136080.387360.166250.161640.279010.353480.189670.296610.254950.34478

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