other errors 0.0028541720000000003
error correction 0.0028343500000000002
spelling errors 0.002829363
error word 0.00282891
training data 0.002628868
grammar errors 0.0026192800000000003
correct errors 0.0026155980000000002
different errors 0.002605188
preposition errors 0.0025536300000000003
grammatical errors 0.002528521
tion errors 0.002481436
common errors 0.0024793980000000003
number errors 0.0024738670000000003
form errors 0.002432882
determiner errors 0.002425958
agreement errors 0.002417137
replacement errors 0.002407294
various errors 0.0023915620000000003
deletion errors 0.002382726
correction system 0.00237826
ment errors 0.002324637
settings errors 0.0023169130000000003
ber errors 0.002298573
matical errors 0.0022894060000000003
placement errors 0.0022856820000000003
grammar error 0.00225626
grammatical error 0.0021655010000000002
test data 0.002097092
errors 0.00205183
error types 0.002044785
testing data 0.002018115
ing data 0.00200666
error cor 0.002003009
development data 0.001972522
ferent error 0.001942185
data selection 0.0019322809999999999
error interaction 0.001926862
error interactions 0.00192356
error inter 0.00192304
word features 0.00190664
data preparation 0.001886058
data sparseness 0.0018836019999999998
correction systems 0.0018519069999999999
correction candidate 0.001851401
candidate word 0.0018459610000000001
correction task 0.0017638950000000001
same word 0.001729506
different correction 0.001698898
error 0.00168881
correction models 0.001682569
language model 0.001631356
correction candidates 0.001611552
possible correction 0.0015906240000000001
final system 0.001590166
word sequence 0.001564488
our system 0.0015531380000000001
rection system 0.0015233780000000001
level correction 0.001510048
other systems 0.001508709
word level 0.001504608
final correction 0.001502986
training set 0.001490856
system outputs 0.0014894420000000001
extra correction 0.00148614
system overview 0.0014771110000000001
correction confidence 0.001475186
system architecture 0.0014704680000000002
nois system 0.001467749
original correction 0.001442274
model score 0.001440921
original word 0.0014368340000000001
word sequences 0.001428918
reasonable correction 0.001424966
context words 0.00141212
language models 0.001404213
individual correction 0.001402193
correction candi 0.001398669
ror correction 0.00139506
jth correction 0.001379334
correction confi 0.001379334
same corpus 0.0013605330000000001
noun phrase 0.001322451
possible words 0.001320459
limited training 0.001297299
original training 0.001280452
english newswire 0.0012771870000000002
correct candidate 0.001269629
training instances 0.0012639399999999999
training instance 0.001263604
other classification 0.001260028
different systems 0.001259725
enough training 0.001255001
english teachers 0.001253699
ing sentences 0.001240775
chosen training 0.001238769
alterative training 0.001235863
special language 0.001234105
system 0.00123272
classification model 0.0012218580000000001
generating training 0.001220674
