error model 0.0034899700000000002
correction model 0.00339309
error correction 0.0030044800000000003
error data 0.00289176
model probability 0.002834826
spelling error 0.00281134
correction data 0.00279488
spelling correction 0.00271446
language model 0.002711764
pinyin error 0.00269999
character error 0.00268154
conversion model 0.002679601
pinyin correction 0.00260311
word string 0.002540499
model approach 0.002539984
word language 0.002537494
model features 0.0025139109999999997
model results 0.0025043319999999997
model models 0.002441577
translation model 0.002436632
model baseline 0.002430847
pinyin spelling 0.00240997
model probabilities 0.00236822
correction probability 0.002349336
model weights 0.002336684
channel model 0.002325204
statistical model 0.002305146
conversion error 0.002290991
smt model 0.002273891
linear model 0.002246345
spelling errors 0.002232758
correction string 0.0022292790000000002
word frequency 0.002224084
trigram model 0.002221468
unified model 0.002210552
integrated model 0.0022099199999999998
model outputs 0.002191244
model adapts 0.002188347
pinyin words 0.0021877800000000003
confusion word 0.00213835
next word 0.002108552
word prediction 0.002057303
word trigram 0.002047198
correction accuracy 0.00204335
word bigram 0.002042671
error pairs 0.002024929
appropriate word 0.002015128
word completion 0.002015128
correction feature 0.001994919
correction research 0.001988881
character edit 0.001986156
online error 0.001982559
training data 0.001969841
correction translation 0.0019511420000000001
such data 0.0019462169999999999
correction baseline 0.001945357
model 0.00193929
data set 0.001937968
correction pairs 0.001928049
pinyin string 0.001924789
error analysis 0.001924102
error rate 0.001907917
correction candidate 0.0019054750000000002
different data 0.001900083
pinyin conversion 0.001889621
correction problem 0.0018737130000000002
relative error 0.001872598
character conversion 0.001871171
paired error 0.0018536
correction task 0.001852413
minimum error 0.001842805
actual error 0.001842534
error patterns 0.0018346220000000002
correction rate 0.0018110370000000002
test data 0.001807934
generating error 0.001799728
driven error 0.001799728
correction candidates 0.001798791
multiple correction 0.001795661
pinyin input 0.001776161
input pinyin 0.001776161
correction module 0.001743118
wrong correction 0.001739749
correction behavior 0.00173316
correction component 0.001731963
correction post 0.0017232850000000002
pinyin characters 0.001715939
data analysis 0.001714502
pinyin sequence 0.0017081800000000001
unseen correction 0.00170559
correction targets 0.001702814
chinese pinyin 0.001696193
online spelling 0.001692539
keystroke data 0.0016765109999999999
correct pinyin 0.001670529
paired data 0.0016439999999999998
chime data 0.001639261
data collection 0.0016376399999999999
conversion probability 0.0016358470000000002
parallel data 0.001622576
