training data 0.00387747
test data 0.003475983
unlabeled data 0.00325267
supervised data 0.003153799
baseline model 0.003081159
web data 0.0030476440000000004
model adaptation 0.0030357
ing data 0.0029957350000000002
supervised model 0.002981049
same model 0.0029534590000000003
model error 0.00291588
wikipedia data 0.002892669
labelled data 0.0028249190000000004
unlabelled data 0.002819334
crf model 0.002813431
data points 0.0028059310000000002
train model 0.002804445
ideal data 0.0028016050000000004
data point 0.0028016050000000004
final model 0.002778695
character word 0.00274037
new model 0.00273886
line model 0.002700931
model expectation 0.002680026
model interpolation 0.002676455
second model 0.002656677
exponential model 0.002653995
model mpl 0.002647306
nal model 0.002643529
model mplea 0.00262937
model 0.00237126
word segmentation 0.0023499330000000002
chinese word 0.002326025
natural word 0.002240007
word sequence 0.002206633
feature function 0.002067302
word seg 0.002037288
word boundary 0.002007467
nese word 0.002004232
word segmenta 0.001957869
word boundaries 0.001957841
training set 0.001954901
word bigram 0.001953811
word unigrams 0.001953405
training corpus 0.001942132
training sentences 0.001917052
label learning 0.001907803
first training 0.001852949
feature set 0.001849201
domain training 0.001834845
crf training 0.0017756310000000002
learning algorithm 0.0017387560000000002
supervised learning 0.0017099290000000002
wikipedia training 0.0016821190000000001
labelled training 0.001614369
training instance 0.001594607
criminative training 0.0015928090000000002
training materials 0.001591192
additional feature 0.001588074
first character 0.001567509
character sequence 0.0015623030000000001
test set 0.001553414
feature func 0.001514924
baseline performance 0.001491912
feature functionφj 0.001486837
feature augmentation 0.001486837
single character 0.001452646
current character 0.0014526
machine learning 0.001414509
supervised performance 0.0013918020000000001
different labels 0.00137763
input features 0.001375776
learning excels 0.0013586240000000001
last character 0.0013494359999999999
consecutive character 0.001348373
statistical features 0.001343914
training 0.00133346
label sequence 0.001321946
sequence label 0.001321946
multiple features 0.0013217910000000001
supervised baseline 0.0013196879999999999
character bigrams 0.00131292
character unigrams 0.001309075
beginning character 0.001309021
above features 0.00130758
wikipedia test 0.001280632
adaptation approach 0.001254106
label distribution 0.001248418
cost function 0.001236005
feature 0.00122776
structural information 0.0011989280000000001
label measurement 0.001177764
domain adaptation 0.001165825
boundary information 0.001161245
baseline system 0.0011518119999999999
large set 0.001144232
valuable information 0.001131151
tual information 0.0011302690000000002
rich information 0.001125018
consistent label 0.001124535
