word features 0.00517491
chinese word 0.003947137
word information 0.003921776
word segmentation 0.003898354
new word 0.0038619680000000003
crf word 0.003849332
word bigram 0.003472853
word recognition 0.0034613400000000002
word detection 0.0034336930000000002
word types 0.0034328180000000002
word list 0.00340917
fast word 0.003364518
word unigram 0.003354146
word identification 0.003342843
word unigrams 0.003341777
word reduplication 0.0033344810000000003
word segments 0.003334063
word segment 0.003333794
training data 0.00312332
different features 0.002894281
new features 0.0028642579999999997
new training 0.002614448
training method 0.002476054
bigram features 0.002475143
training algorithm 0.0024287930000000003
dimensional features 0.00239903
edge features 0.002394699
sparse features 0.002360766
training set 0.00235972
baseline training 0.002345892
refined features 0.0023447389999999998
node features 0.002338725
edges features 0.0023357399999999998
perceptron training 0.0023030760000000003
training methods 0.0023021310000000002
training process 0.0022969180000000002
training time 0.002277126
new words 0.002259998
online training 0.00223856
training sample 0.002216023
training speed 0.0021607930000000003
sgd training 0.002156514
msr training 0.002136835
batch training 0.002131249
training the 0.0021249140000000003
other words 0.002119932
fast training 0.002116998
pku training 0.002111285
adf training 0.0021096080000000002
popular training 0.002092105
lbfgs training 0.002091071
training samples 0.002089505
features 0.0020886
accelerated training 0.002086396
training cost 0.002085417
feature vector 0.002030712
segmentation model 0.002027434
feature templates 0.001991121
crf model 0.001978412
feature function 0.001930094
frequency feature 0.001922774
feature frequency 0.001922774
data method 0.001921794
output words 0.001880059
chinese character 0.001840881
training 0.00183879
feature dimension 0.001833848
edge feature 0.001831369
feature sets 0.001816699
global feature 0.001790027
short words 0.001732167
chinese information 0.001696293
corresponding data 0.001655888
markov model 0.001614737
joint model 0.001583756
msr data 0.0015825750000000001
probabilistic model 0.001564503
pku data 0.001557025
learning method 0.0015516689999999999
model parameters 0.001528879
feature 0.00152527
cws model 0.001514232
learning algorithm 0.001504408
character sequence 0.001499002
words 0.00148434
label set 0.001453649
label sequence 0.001451667
learning rate 0.001399959
different values 0.001398918
learning methods 0.001377746
chinese text 0.001371681
small learning 0.001348685
character types 0.0013265619999999999
online learning 0.001314175
local character 0.001312953
possible label 0.0012982410000000001
gradient information 0.001294434
segmentation results 0.001291006
segmentation system 0.0012610850000000001
single label 0.001260403
