model test 0.003487228
model models 0.00342735
model segmentation 0.003267331
segmentation model 0.003267331
model learning 0.003226085
crf model 0.003065834
markov model 0.003017638
baseline model 0.0029476380000000003
tagging model 0.002913152
previous model 0.002889584
entropy model 0.002869725
model interactions 0.002864944
labeling model 0.002860435
basic model 0.002843254
sensitive model 0.002797594
powerful model 0.002794844
represenation model 0.002789953
chine model 0.002786718
model 0.00255564
chinese word 0.002526787
word segmentation 0.0025091510000000003
domain word 0.0024623730000000003
training data 0.00244825
test data 0.002226828
feature set 0.002205922
oov word 0.0021976
word recall 0.002135951
traditional word 0.002088764
word count 0.002083353
forms word 0.002071267
word seg 0.002064146
type features 0.002052728
word break 0.002033612
nese word 0.002029438
word breaks 0.002026894
input features 0.0020266710000000003
new feature 0.002006258
same feature 0.001991165
feature value 0.0019746
training set 0.001913832
bigram feature 0.001884752
feature templates 0.001868415
extract features 0.0018650390000000002
type feature 0.0018632179999999998
different test 0.001786785
feature representation 0.001777949
basic feature 0.001732714
feature representations 0.001725711
new training 0.001714168
above feature 0.001701319
computation feature 0.0016994319999999999
test set 0.00169241
set test 0.00169241
whole feature 0.001690519
combined feature 0.0016865909999999999
feature template 0.001681007
feature tem 0.001680158
feature represenation 0.0016794129999999998
represenation feature 0.0016794129999999998
feature representa 0.001677532
training error 0.001670094
features 0.00163461
chinese words 0.001621038
pku data 0.0016116540000000001
training problem 0.001586598
segmentation models 0.001583401
learning models 0.001542155
training texts 0.001474164
pku training 0.001469424
segmented training 0.0014691110000000002
tag set 0.001446192
feature 0.0014451
chinese segmentation 0.001441018
new models 0.001432868
set domain 0.001425735
segmentation segmentation 0.001423382
segmentation results 0.001419893
training domains 0.0014139640000000002
chinese character 0.001404247
enlarged training 0.0013871310000000002
training materials 0.0013832890000000002
training principles 0.001382346
chinese language 0.001353903
chinese information 0.001350494
test methods 0.00133128
oov words 0.0012918510000000001
other domain 0.001286719
segmentation performance 0.0012595
test sets 0.001252691
second test 0.001246092
third test 0.001244755
ing set 0.001230106
segmentation approach 0.00122967
label sequence 0.0012288009999999999
closed test 0.0012237580000000001
english words 0.0012206
statistics test 0.0012139849999999999
same domain 0.0012109780000000001
hypothesis test 0.00120987
test domains 0.001192542
