pos model 0.00334864
probability model 0.003004357
model distribution 0.002791918
above model 0.0026889060000000005
model probabilities 0.0026707930000000003
order model 0.0026538940000000004
entropy model 0.002578309
model symbol 0.0025444430000000004
trigram model 0.002543893
model distributions 0.0025394500000000004
direct model 0.002527299
generative model 0.00250286
membership model 0.0025004460000000004
markov model 0.002456967
model 0.00215975
training data 0.001922581
other words 0.001842968
pos corpus 0.001830251
word error 0.001808236
test data 0.001788146
pos tag 0.001770644
pos tags 0.001749778
pos tagging 0.0016909999999999998
pos string 0.0016332859999999999
english words 0.001612876
possible pos 0.001573498
pos trigram 0.001573033
probability models 0.001555885
single pos 0.0015469589999999999
training corpus 0.001527832
pos train 0.0015182289999999998
pos symbols 0.0015158629999999999
likely pos 0.00148919
probability distribution 0.001476775
true words 0.001467999
annotated pos 0.0014644839999999998
probable pos 0.001463888
pos trigrams 0.001460027
german words 0.00145903
word recognition 0.0014529809999999999
other decision 0.001443241
word membership 0.001442896
different tokens 0.0014257380000000002
word vocabulary 0.00142143
sentence probability 0.0014112270000000001
machine translation 0.0013550699999999999
task decision 0.001326371
decision rule 0.001320435
first error 0.001297091
sentence error 0.001272656
language processing 0.001249445
statistical decision 0.001241048
language modelling 0.0012406309999999999
training criterion 0.001225135
string decision 0.001222559
training procedure 0.001212921
joint distribution 0.001211322
tagging error 0.0012081459999999998
corresponding algorithm 0.001205415
natural language 0.001196658
labelled training 0.001187813
words 0.00117789
cost function 0.001174468
training criteria 0.001173261
marginal probability 0.001168896
posterior probability 0.00116632
experimental results 0.001165502
bayes decision 0.001163593
symbol decision 0.001162856
conventional training 0.0011592920000000001
training procedures 0.0011592920000000001
programming algorithm 0.001158037
string error 0.0011504319999999998
viterbi algorithm 0.0011354730000000001
efficient algorithm 0.001131495
entropy models 0.0011298369999999999
specific form 0.001123001
forward algorithm 0.001122522
gis algorithm 0.001122522
error rate 0.0011221019999999998
typical decision 0.00112182
tagging results 0.001118065
test sentences 0.001116276
decision rules 0.001114054
similar results 0.0011120549999999998
tag sequence 0.001109357
decision overall 0.001105773
wsj corpus 0.001099145
loss function 0.001091406
symbol error 0.001090729
minimum error 0.00108961
english corpus 0.001076347
tagging task 0.001050318
decision theory 0.0010491300000000001
state sequence 0.001048907
feature parameters 0.0010429319999999999
error measure 0.0010322929999999999
tag string 0.00102615
trigram distribution 0.001016311
bleu measure 0.001010795
