character word 0.0052743600000000005
word features 0.0042797
word segmentation 0.004170504
new word 0.003911739
first word 0.003828218
word information 0.0038204809999999997
joint word 0.0037789539999999997
word length 0.003777504
character words 0.0037604
word candidate 0.003710715
partial word 0.003624221
maximum word 0.003588328
second word 0.003586498
last word 0.003584721
single word 0.003580879
word seg 0.0035703849999999997
current word 0.003566474
full word 0.003551519
complete word 0.00351224
frequent word 0.003509497
vious word 0.003491898
word frequencies 0.003490697
pure word 0.0034872659999999997
tial word 0.0034871019999999997
word informa 0.003483697
plete word 0.003480072
quent word 0.0034799569999999997
words training 0.00288533
character sequence 0.0028361610000000002
chinese character 0.0028304050000000002
first character 0.0026652380000000003
character type 0.0025236390000000003
previous character 0.002474731
next character 0.002451275
second character 0.0024235180000000004
last character 0.0024217410000000003
current character 0.0024034940000000004
final character 0.002382742
initial character 0.0023741150000000004
other words 0.0023413369999999998
third character 0.0023403020000000003
incoming character 0.002322194
ith character 0.0023221820000000003
starting character 0.002320481
rent character 0.002319747
unknown words 0.0023105549999999997
previous words 0.002123751
sentences words 0.0021195289999999998
partial words 0.002110261
segmented words 0.002060037
character 0.00205569
full words 0.002037559
complete words 0.00199828
frequent words 0.001995537
known words 0.001975595
mented words 0.001974171
sible words 0.0019657249999999998
model approach 0.001920392
perceptron training 0.001815816
training data 0.001812482
pos tags 0.001785215
same training 0.0017845640000000002
training algorithm 0.0017798570000000001
first pos 0.001729468
linear model 0.001722044
pos information 0.0017217309999999998
words 0.00170471
training case 0.0017038560000000001
segmentation error 0.001690171
segmentation tags 0.0016171290000000001
possible pos 0.001596667
single model 0.0015903290000000001
model scores 0.001588991
training example 0.001565748
training process 0.0015615350000000002
final model 0.001555172
training iteration 0.001539032
different approach 0.0015225949999999999
global model 0.0015178650000000002
joint segmentation 0.001512118
development training 0.001507309
ensemble model 0.001489049
training examples 0.001473355
partial features 0.0014665809999999998
training iterations 0.0014657630000000001
incremental training 0.001458586
consistent training 0.001444398
such decoding 0.00142845
feature templates 0.001423166
pos prediction 0.0014071169999999998
such char 0.0013925370000000001
pos category 0.001391815
pos enumeration 0.001380498
feature value 0.001379076
such comparison 0.001366702
other characters 0.001357404
feature vector 0.001349107
lookahead features 0.001347879
chinese tree 0.001301751
feature template 0.001298042
