word segmentation 0.0032225220000000002
word classes 0.003072242
word string 0.003054756
word sequence 0.002956675
language model 0.00288216
word boundaries 0.002862211
word history 0.002841417
each word 0.002809159
supposed word 0.002805862
chinese words 0.002682837
model models 0.002542729
english words 0.002314751
oov words 0.002312165
unknown words 0.002284352
words sequence 0.002256235
optimal words 0.00221767
clustering words 0.002210259
model level 0.002133468
various model 0.002132569
model adaptation 0.002122981
words entries 0.002111812
model unit 0.002069315
model units 0.002062954
markov model 0.002062343
model groups 0.002033251
proper model 0.002017141
guage model 0.002013752
pos language 0.001993343
model topology 0.001991101
language models 0.001913549
speech data 0.00190611
words 0.00186959
other data 0.001829817
segmentation language 0.001778982
model 0.00175567
training data 0.001733675
news data 0.001654125
pos information 0.001542927
segmentation pos 0.001519345
test corpus 0.001510868
speech information 0.0015082440000000002
bigram language 0.0015027039999999999
speech training 0.001491905
evaluation data 0.001458801
speech recognition 0.001458717
natural language 0.001443467
adaptation data 0.001441251
language modeling 0.001440184
language scores 0.001428679
gram language 0.001421419
different tags 0.001400644
western language 0.001399191
language mod 0.0013911169999999999
language processing 0.001386878
data ing 0.001383706
chinese names 0.001379998
structured language 0.001362084
superarv language 0.001362084
different level 0.0013546839999999999
years data 0.001348681
june data 0.001311474
december data 0.001309907
may data 0.001309907
uation data 0.001309907
adaption data 0.001309907
data sparsity 0.001309907
pos tags 0.001290611
chinese character 0.0012873770000000001
chinese oov 0.001255822
pos sequence 0.001253498
acoustic models 0.001253451
recognition system 0.001242844
other methods 0.0012354879999999999
different levels 0.0012141259999999998
different styles 0.0012141259999999998
baseline speech 0.001206067
test set 0.001200508
names recognition 0.001193298
segmentation results 0.0011787619999999999
first character 0.0011757500000000001
pos tagging 0.001168507
speech corpora 0.00116343
linguistic knowledge 0.001160895
analyzer models 0.0011560070000000001
recognition results 0.001152817
pos wfsts 0.001148154
new knowledge 0.0011423000000000002
continuous speech 0.001137809
chinese characters 0.001136454
training set 0.0011321670000000001
language 0.00112649
male speech 0.001125699
toy pos 0.0011254519999999999
models step 0.001122849
mandarin speech 0.00111932
name recognition 0.001119167
acoustic segmentation 0.001118884
markov models 0.001093732
name class 0.001084658
recognition systems 0.001073099
