word form 0.003903186
joint word 0.0038926969999999997
word segmentation 0.0037138429999999997
word boundaries 0.0036258899999999997
japanese word 0.003611071
word lattice 0.003574366
word recognition 0.003533981
word seg 0.003505801
word surface 0.0035026939999999998
pos tag 0.0034785199999999997
modeling word 0.003467964
only word 0.003466546
term word 0.0034538579999999998
accurate word 0.003448437
word derivation 0.00343789
word segmenta 0.003432081
word segmen 0.0034319109999999997
word lat 0.0034258409999999998
hance word 0.003424851
word lengthening 0.003424851
model features 0.0029405
pos tags 0.002894742
pos tagging 0.0026184529999999998
joint model 0.002524117
language model 0.002514504
model feature 0.002489233
normal pos 0.00243997
surface pos 0.0023951239999999998
juman pos 0.002365789
tagging model 0.002357443
mal pos 0.002336921
distinct pos 0.002326794
face pos 0.002322059
training corpus 0.00231696
pos norm 0.002316956
such words 0.002312924
current model 0.002181686
model this 0.0021519069999999998
corpus form 0.002127826
statistical model 0.002123794
baseline model 0.002108488
discriminative model 0.002081448
accurate model 0.002079857
inative model 0.002057364
single words 0.001954103
tag set 0.0019406570000000001
tag dictionary 0.001929979
corpus analysis 0.001926538
informal words 0.001831338
vocabulary words 0.001795945
microblog corpus 0.001779968
model 0.0017691
annotated corpus 0.001758033
corpus this 0.001745127
tag dictionaries 0.001738579
tag dictionar 0.001736177
sms corpus 0.00171144
corpus annotation 0.00170141
chat corpus 0.001672705
raw corpus 0.001662582
croblog corpus 0.001657688
notated corpus 0.001649904
binary features 0.001608961
new features 0.001585833
original features 0.001585093
joint models 0.001521334
training method 0.001508667
words 0.00148512
sophisticated features 0.001461641
such rules 0.001435974
system performance 0.001377357
corpus 0.00136232
training methods 0.0013431139999999999
perceptron training 0.001324899
training example 0.001301727
joint task 0.001267954
further performance 0.001259935
first time 0.001258639
plored performance 0.00118294
perior performance 0.00118294
performance sig 0.00118294
performance gains 0.00118294
performance gain 0.00118294
such corpora 0.001182051
current models 0.001178903
vowel character 0.001178431
normal form 0.001175366
features 0.0011714
binary feature 0.001157694
hybrid approach 0.001156144
programming algorithm 0.0011438680000000001
rich information 0.001132133
surface form 0.0011305199999999999
form news 0.00113025
statistical models 0.0011210109999999999
such actions 0.0011174380000000001
experimental results 0.001112062
standard rules 0.001111957
error analysis 0.001111935
correct analysis 0.001096465
