word segmentation 0.003028129
word spacing 0.003000358
new word 0.0029536469999999998
correct word 0.002875029
proper word 0.0028691019999999997
word boundary 0.002833601
baseline word 0.002817292
average word 0.002811521
word boundaries 0.002752096
initial word 0.002722712
word spac 0.002699142
segmentation model 0.0021030989999999998
training data 0.0020941880000000003
markov model 0.001968281
input method 0.001932447
test data 0.0019318970000000001
japanese language 0.001929987
model baseline 0.001892262
korean language 0.001817781
model proposed 0.0017788399999999998
development data 0.0017548630000000002
method web 0.001690551
ment data 0.001656438
training corpus 0.001605103
evaluation method 0.001541956
model 0.00149724
user input 0.0014953000000000002
spacing information 0.001493832
method figure 0.001477121
test corpus 0.001442812
language 0.00142377
correction method 0.001418467
proposed method 0.00139301
method outper 0.001387974
spacing error 0.001307342
different set 0.0013046070000000002
threshold function 0.001263898
corpus lee 0.0012316
similar error 0.001226333
same output 0.001208286
quency information 0.001197227
error rate 0.001195007
text web 0.001169052
web text 0.001169052
etri corpus 0.001167233
sejong corpus 0.001167233
proper training 0.00115571
human error 0.001143817
spacing errors 0.001136963
method 0.00111141
new spacing 0.001109465
other nlp 0.00110678
input ini 0.001090855
spelling errors 0.001089825
ing error 0.001082505
error patterns 0.001075851
other languages 0.001039352
estimation error 0.00102346
error range 0.001002138
original user 9.99833E-4
usual error 9.990420000000001E-4
spacing states 9.89145E-4
same way 9.86846E-4
spacing state 9.49258E-4
first step 9.459410000000001E-4
test result 9.354719999999999E-4
traditional approach 9.29327E-4
simple threshold 9.281720000000001E-4
web texts 9.2568E-4
small number 9.25197E-4
experimental results 9.22483E-4
possible combinations 9.20908E-4
information 9.15744E-4
ing errors 9.121260000000001E-4
hidden markov 9.11708E-4
spacing quality 9.00238E-4
corpus 8.96225E-4
first experiment 8.863950000000001E-4
research issue 8.8466E-4
few spacing 8.81745E-4
performance improvement 8.80578E-4
enough test 8.754839999999999E-4
threshold estimation 8.7365E-4
important contribution 8.672840000000001E-4
nlp applications 8.61267E-4
variable threshold 8.597030000000001E-4
web portals 8.523160000000001E-4
ter output 8.488710000000001E-4
further nlp 8.44311E-4
korean sentences 8.42754E-4
spelling correction 8.38007E-4
such wbms 8.32126E-4
recent research 8.32092E-4
output optimization 8.27957E-4
practical nlp 8.262250000000001E-4
many researchers 8.24552E-4
able output 8.23385E-4
input 8.21037E-4
pseudo test 8.16813E-4
research subject 8.074460000000001E-4
