From: Applying deep matching networks to Chinese medical question answering: a study and a dataset
Input Unit
Model
P@1(%)
MAP(%)
Random
20.0
45.7
Char
BM25
26.6
51.2
multiCNN[17]
39.8
60.1
MV-LSTM
58.1
74.5
MatchPyramid
66.0
79.3
Word
23.6
49.0
40.0
60.5
59.3
75.3
58.8
74.9