unknown word 0.002839899
word accuracy 0.00270052
word tags 0.002686446
unknown words 0.002649369
other words 0.002613519
word rules 0.002359504
many words 0.0023488100000000002
times word 0.002284422
word distributions 0.002267901
tagging model 0.002260761
word class 0.002251229
known word 0.002245773
word endings 0.002235744
word accu 0.002215654
statistical model 0.002213091
markov model 0.002172897
lexicon words 0.002163432
new model 0.002115618
trigram model 0.002107742
known words 0.002055243
digit words 0.002054483
capitalized words 0.002023253
tag sequence 0.001984407
separator model 0.00194765
possible tag 0.0019427770000000001
trigram tag 0.0017634120000000002
words 0.0017554
training data 0.0017052740000000001
special tag 0.001703246
bigram tag 0.00169739
tagger data 0.001684765
ing data 0.001666286
probability distribution 0.001655129
model 0.00165354
tag symbols 0.0016313240000000002
tag triples 0.001629255
entire tag 0.001628914
likely tag 0.001606976
data estimation 0.001588444
statistical data 0.001576231
ski tag 0.0015749310000000001
training corpus 0.001554916
corpus tagger 0.001534407
smoothing method 0.001496727
new method 0.001459278
test set 0.00144608
feature information 0.001441053
new probability 0.001386089
state distribution 0.0013832900000000001
trigram probability 0.001378213
unknown overall 0.0013732829999999999
unknown standard 0.001372492
hmm tagger 0.001371124
lexical information 0.0013650329999999999
testing data 0.001364163
data problem 0.001359579
testing method 0.001344683
common method 0.001344146
data problems 0.001335908
sparse data 0.001328566
state sequence 0.0013273690000000001
other systems 0.001315783
transition probability 0.001269707
ditive method 0.001265123
probability esti 0.001253898
wsj corpus 0.001250807
probability distributions 0.001245982
lexical probabilities 0.001237555
overall accuracy 0.001233904
initial state 0.0012266339999999999
markov models 0.001226011
valid probability 0.001216374
other taggers 0.001211934
probability distri 0.0011985
journal corpus 0.001195168
probability aij 0.00119465
probability aijk 0.001190509
standard hmm 0.0011815620000000002
suffix information 0.0011768730000000002
brown corpus 0.001166203
current models 0.001138483
contextual information 0.001138092
speech tagging 0.001125857
other researcher 0.001125292
trigram tagger 0.001122287
new state 0.00111425
different length 0.001109185
trigram probabilities 0.001100936
additional feature 0.001097949
accuracy rate 0.0010969859999999999
ing approach 0.001093641
specific information 0.001091547
bigram hmm 0.001091219
lexical smoothing 0.001090348
hmm system 0.001086277
current state 0.001084001
accuracy levels 0.001072893
general distribution 0.001072885
new tagging 0.0010692990000000001
tagger type 0.001059034
