word errors 0.00317454
different errors 0.00279104
error corpus 0.00258951
error correction 0.00255541
spelling errors 0.002321534
spelling error 0.002187114
many errors 0.002174327
typing errors 0.0020156279999999998
correct word 0.002012241
real errors 0.002000324
error corpora 0.001994871
artificial errors 0.001993572
natural errors 0.001985854
error models 0.0019856730000000003
main error 0.001984295
text data 0.0019739939999999997
genuine errors 0.001967895
errors changes 0.001926735
error source 0.0019199500000000001
correction system 0.0018755290000000001
real error 0.001865904
artificial error 0.0018591520000000002
similar words 0.0018347889999999999
error sources 0.001815269
foreign words 0.001814595
erroneous word 0.001810558
long word 0.0017720700000000002
ambiguous word 0.001751334
intended word 0.001747807
word trigram 0.001736236
unknown word 0.001724099
unrecognised word 0.001720429
real words 0.001693614
frequent words 0.001690022
errors 0.00168586
few words 0.001684905
english corpus 0.0016772900000000001
short words 0.001668525
long words 0.0016625400000000001
spelling correction 0.001639644
consecutive words 0.0016242589999999999
unknown words 0.001614569
compound words 0.001613312
error 0.00155144
different corpora 0.001548611
correction systems 0.00154585
different proposals 0.001542537
correction candidate 0.00152228
frequency data 0.001505176
other context 0.001488839
different techniques 0.001475002
correction proposals 0.001441327
different factors 0.0014258790000000001
other corpora 0.001425626
automatic correction 0.001416394
single correction 0.0013870520000000002
words 0.00137915
correction techniques 0.001373792
different ways 0.0013550230000000001
different questions 0.001352993
different methods 0.001351362
language models 0.001349863
different kinds 0.0013475260000000001
artificial corpus 0.0013457820000000002
different alternatives 0.0013418570000000001
context features 0.001324091
same text 0.0013227360000000001
brown corpus 0.001306649
seed corpus 0.001269897
selecting corpus 0.001269897
correction proposal 0.001268623
sensitive correction 0.001242627
following features 0.001235249
text sentences 0.001163342
present system 0.001151136
language mismatch 0.001146188
real text 0.001117818
training corpora 0.001113554
same sentence 0.001109924
first results 0.0010893040000000001
magazine text 0.0010775770000000001
raw text 0.001064045
automatic spelling 0.001048098
corpus 0.00103807
french text 0.001037766
uppercase character 0.001030687
candidate corrections 0.0010214920000000001
combinations table 0.0010076759999999999
correction 0.00100397
possible spelling 9.94582E-4
multiple table 9.85309E-4
training purposes 9.492400000000001E-4
simple method 9.413E-4
many proposals 9.25824E-4
english texts 9.21027E-4
genuine spelling 9.177090000000001E-4
language 9.1563E-4
good results 9.15009E-4
automatic evaluation 9.14375E-4
noun corrections 9.130060000000001E-4
