error sentences 0.0035899
error correction 0.003163247
error distribution 0.003033221
english error 0.002975365
grammar error 0.002911614
large error 0.002897015
error generation 0.002889582
particle error 0.00283635
error patterns 0.0028160620000000003
error cor 0.002792096
error pat 0.002788937
error corpora 0.002780505
error types 0.002772521
error distribu 0.0027405420000000003
error gen 0.002739949
error corrector 0.002739308
error correc 0.002739308
error genera 0.002739308
error 0.00245092
target feature 0.002419116
word sequence 0.002372535
feature space 0.002272427
feature weight 0.002241044
binary features 0.002191628
language model 0.002189035
output word 0.002172288
review feature 0.002153332
feature weights 0.0021432309999999998
common features 0.002140762
feature augmentation 0.002129993
feature aug 0.002129649
word alignment 0.00212805
above feature 0.0021254439999999998
entire feature 0.002122526
domain data 0.00205826
real features 0.002043201
mapping features 0.002027077
link features 0.002023163
gram features 0.002008934
actual word 0.001982163
word sequences 0.001963785
word order 0.001934425
empty word 0.001914194
model probability 0.001909228
training corpus 0.001813636
feature 0.00181248
other errors 0.0017216700000000002
features 0.00170926
input sentences 0.001700568
new sentences 0.001691479
correct sentences 0.001687865
output sentences 0.001686748
target data 0.001682596
discriminative model 0.0016660820000000002
target domain 0.001588936
input words 0.001573828
trigram model 0.001572044
output words 0.0015600079999999999
sequence conversion 0.001551831
main data 0.001496814
conversion approach 0.001490923
paired sentences 0.001490639
source domain 0.001472105
domain adaptation 0.001443713
data figure 0.001426774
correction method 0.001422506
supervised data 0.001395673
particle errors 0.001378122
unchanged words 0.001348903
additional corpus 0.001318222
surrounding words 0.001310629
communication errors 0.0013072350000000001
chine translation 0.001297434
system correction 0.0012972
tiple errors 0.0012816630000000002
corpus con 0.001270005
language analyzer 0.001255142
phrase sequence 0.001248068
correction task 0.001231717
model 0.00123057
large sentence 0.0011905589999999999
discriminative sequence 0.001183527
training corpora 0.001176459
sentences 0.00113898
following sentence 0.001119941
generation probability 0.00111732
source target 0.001096441
morpheme conversion 0.0010940379999999999
results figure 0.001082639
tion probability 0.001068652
target space 0.001066583
similar approach 0.001061982
sequence con 0.001051258
output phrase 0.001047821
tive sequence 0.001047102
rect sentence 0.0010407
such particle 0.001034432
ment results 0.001033973
other hand 0.001025901
system performance 0.00102029
