opinion sentences 0.0034640799999999996
such sentences 0.00310489
negative sentences 0.002862175
opinion sentence 0.0028529799999999998
opinion classification 0.00260401
positive sentences 0.0025545719999999997
sentence classification 0.0025091899999999997
several sentences 0.0024956469999999997
japanese sentences 0.002393971
negative opinion 0.0023458949999999998
original sentences 0.00233076
individual sentences 0.0023052359999999996
tive sentences 0.002286026
ion sentences 0.0022831469999999997
negative sentence 0.002251075
sentiment classification 0.002063793
other words 0.002019184
sentences 0.00199018
opinion sen 0.001948534
positive sentence 0.001943472
extraction opinion 0.001927094
other data 0.001895974
sentence extraction 0.001832274
press opinion 0.001831874
sentence level 0.001822877
label opinion 0.00180478
above sentence 0.001789223
japanese sentence 0.0017828710000000001
extracts opinion 0.001772758
single sentence 0.001767882
entire sentence 0.001760221
sentence cost 0.001712408
sentence classifica 0.001689585
sentence figure 0.001683485
sentence boundary 0.001665894
first method 0.001643023
much information 0.001629994
such phrases 0.0016207860000000001
subjective words 0.001613369
such sen 0.0015893440000000001
classification next 0.001511136
such texts 0.0014837270000000002
opinion 0.0014739
synonymous words 0.00145748
data sets 0.001456167
specific corpus 0.001451708
such item 0.0014381700000000001
such items 0.0014303610000000002
classification result 0.001429241
timent classification 0.0014291599999999998
corpus con 0.001425128
classification turney 0.001418206
such applications 0.001415568
such itemization 0.001413729
automatic corpus 0.001413398
such cell 0.00140634
corpus construction 0.001398868
sentence 0.00137908
corpus crossvalidation 0.001367882
other domain 0.0013415369999999999
second method 0.0013385419999999999
japanese word 0.001334078
other tags 0.001327547
dependency structure 0.001326613
word order 0.001316852
corpus design 0.001316329
dependency pattern 0.0013155929999999999
similar method 0.001303428
negative expressions 0.0012916989999999999
japanese dependency 0.0012696019999999999
unsupervised sentiment 0.0012554110000000001
negative precision 0.001255213
different domains 0.001253791
bootstrapping method 0.00125024
different precision 0.00124538
precise model 0.0012447
noise text 0.001241429
automatic method 0.001236446
negative content 0.001227685
construction method 0.001221916
other labels 0.001202242
negative ones 0.001199957
negative computer 0.001192296
third method 0.001185373
negative indicator 0.001180554
words 0.00117009
negative indicators 0.001169972
first idea 0.001169702
first column 0.001168311
unsupervised method 0.001167796
dependency representations 0.001165765
other works 0.001165315
novel method 0.001164298
information 0.00115986
other researchers 0.001156171
dependency rep 0.001155323
html documents 0.001153387
other languages 0.0011480470000000001
other hand 0.001146944
method extracts 0.0011449259999999999
