word model 0.0027889200000000003
other features 0.002689177
dependency features 0.002612216
reranking features 0.00254873
feature set 0.0023953940000000003
additional features 0.002388214
word probability 0.002345522
strong features 0.002301691
morphological features 0.002300408
various features 0.002290014
binary features 0.002287488
useful features 0.002279712
related features 0.002276925
distance features 0.002263475
dependent features 0.002241646
dency features 0.002235478
gram features 0.002235121
pendency features 0.002231856
phological features 0.002223926
informative features 0.00222247
affixation features 0.002220668
morph features 0.002220668
affixpos features 0.002220668
training data 0.0021155279999999998
word sequence 0.002100432
new feature 0.0020876460000000003
unknown word 0.002037454
features 0.00199036
word pairs 0.00197884
word emission 0.001970766
word end 0.001968494
chinese word 0.001966887
feature types 0.0019366070000000001
binary feature 0.0019349880000000001
word class 0.001928271
feature weights 0.0019273290000000002
test data 0.0019140049999999999
entire word 0.001899303
combined feature 0.001890905
word formation 0.001887521
word help 0.001885766
word endings 0.001880266
ing data 0.0017589489999999999
reranking model 0.00169711
hmm model 0.0016468070000000001
feature 0.00163786
baseline model 0.001625623
pos tag 0.001622257
pos tagging 0.001609718
data the 0.0016014039999999998
training set 0.0015928119999999999
development data 0.001580333
data files 0.001510022
model chinese 0.0014554470000000001
pos information 0.0014349369999999998
pos reranking 0.0014309639999999998
training algorithm 0.0014261769999999998
probabilistic model 0.001417431
unknown words 0.001405934
test set 0.001391289
training method 0.001388579
mixture model 0.0013687270000000001
english words 0.001353045
chinese words 0.001335367
new pos 0.00132238
pos tags 0.001309054
tag accuracy 0.00129316
tagging accuracy 0.001280621
training methods 0.001277354
inflected words 0.001263015
known words 0.001258745
hmm tag 0.0012577299999999999
training process 0.001248758
hmm tagging 0.001245191
ing set 0.001236233
supervised training 0.001218804
tagging performance 0.0012090410000000001
english pos 0.001206979
same method 0.001202638
smoothing parameter 0.001200782
tag sequence 0.001199915
dependency tree 0.0011978079999999999
pos type 0.001179184
hmm tagger 0.001169083
original pos 0.001167023
training loss 0.0011618359999999999
training error 0.0011563189999999998
tagging problem 0.0011531010000000001
mandarin pos 0.001151376
reranking algorithm 0.001149269
model 0.00113874
weight function 0.001132823
original training 0.001129707
smoothing method 0.001127158
training examples 0.001124772
pos assignment 0.0011229859999999999
first test 0.001122062
previous tag 0.001120665
ing parameter 0.001105624
darin pos 0.001102319
