word error 0.004403880000000001
character model 0.004275650000000001
dictionary word 0.003683355
word stem 0.003605695
stem word 0.003605695
error model 0.00358966
word sequence 0.00358558
correct word 0.0035788910000000003
level word 0.003565073
word level 0.003565073
word corrections 0.003521929
candidate word 0.003482849
word frequency 0.003464437
word trigram 0.003458553
character error 0.0033890500000000002
word prediction 0.0033871080000000002
important word 0.003386858
word surface 0.003378841
language model 0.00337026
clean word 0.003353327
valid word 0.003336954
word clustering 0.003331819
corrupted word 0.0033252250000000002
word boundaries 0.003324296
word collocations 0.003322673
misrecognized word 0.0033212090000000003
word peece 0.0033212090000000003
word elongations 0.0033212090000000003
correction model 0.0032428500000000002
ocr character 0.00309316
character correction 0.00304224
level model 0.002750853
models model 0.002744551
channel model 0.002624693
based model 0.002619324
single character 0.002592046
confusion model 0.002573936
character segment 0.002554119
character level 0.002550243
acter model 0.0025323200000000002
guage model 0.002517121
base model 0.002508969
trained model 0.0025085380000000003
model account 0.0025067320000000002
character recognition 0.002420705
based character 0.002418714
ocr error 0.00240717
character substitution 0.002401741
arabic words 0.00239253
improved character 0.002391387
character segments 0.002391278
character alignment 0.002369144
character substitutions 0.002365433
character image 0.002362031
character position 0.0023606860000000003
error correction 0.00235625
character images 0.0023340270000000002
character mapping 0.002318539
optical character 0.002311322
null character 0.0023111240000000003
character posi 0.0023088600000000003
character substiu 0.0023068800000000003
character codes 0.0023068800000000003
model 0.00223813
ocr errors 0.00212762
possible words 0.0020794050000000003
character 0.00203752
candidate words 0.001952109
arabic ocr 0.00192656
stem error 0.001904875
spelling errors 0.001892286
spelling correction 0.001825026
valid words 0.001806214
segmented words 0.0017980930000000002
consecutive words 0.001795212
rupted words 0.001791024
testing words 0.001791024
words ver 0.001791024
text errors 0.001741648
ocr text 0.001725308
improved error 0.001705397
error cor 0.0017018900000000002
error rate 0.00166807
ocr system 0.001667507
correction training 0.0016589579999999999
arabic information 0.0016360609999999999
error correc 0.0016236690000000001
error rates 0.0016212980000000002
english language 0.001578544
ocr output 0.001543482
arabic text 0.0015405879999999999
trigram language 0.001538333
language modeling 0.0015381980000000002
different edit 0.001534713
words 0.00152161
segment correction 0.001521319
edit distance 0.0014922199999999998
rich language 0.0014728900000000001
recognition errors 0.001455165
complex language 0.001453349
