language model 0.00308899
phrase model 0.002623799
class model 0.0025816750000000003
word bigram 0.002514811
word trigram 0.0024927160000000003
word sequence 0.00246105
word tokens 0.0023663910000000002
bigram model 0.002362471
trigram model 0.002340376
conventional word 0.002292875
word ngram 0.002281359
word prediction 0.002269408
tional word 0.002261751
stochastic model 0.002183209
model size 0.002165917
ngram model 0.0021290190000000002
guage model 0.002128504
multigram model 0.002119411
model nmch 0.002106731
language data 0.002044715
language models 0.001984408
model 0.00188188
language modeling 0.001664631
hesitation words 0.001631892
phrase models 0.001519217
language dis 0.001510187
stochastic language 0.0015084389999999999
cal language 0.001486229
phrase distribution 0.001476579
language mod 0.0014520549999999998
language understanding 0.001437813
class distribution 0.001434455
training data 0.00141732
words 0.00136596
bigram probability 0.001359914
phrase clustering 0.0012618859999999998
bigram models 0.001257889
test data 0.0012529729999999999
trigram models 0.0012357940000000001
conditional probability 0.0012355439999999999
unigram probability 0.001210706
language 0.00120711
class bigram 0.001180386
evaluation data 0.0011351249999999998
unknown data 0.001076511
arrangement data 0.001070503
rangement data 0.001064308
clustering algorithm 0.001064218
incomplete data 0.001062764
same class 0.0010566780000000001
gram models 0.0010533180000000001
class estimates 0.001051565
phrase dis 0.0010449959999999999
class label 0.001044692
stochastic phrase 0.0010432479999999999
training corpus 0.001032701
based models 0.001030479
different length 0.001030128
trigrams class 0.001027434
ngram models 0.001024437
sentence likelihood 0.001019263
bigram perplexity 0.001006739
class assignment 0.001001121
estimation procedure 9.88047E-4
likelihood values 9.79607E-4
phrase retrieval 9.74618E-4
phrase abe 9.678429999999999E-4
common class 9.646780000000001E-4
class versions 9.61528E-4
ciass distribution 9.606220000000001E-4
class version 9.56515E-4
class assumption 9.55323E-4
previous clustering 9.47582E-4
clustering procedure 9.471359999999999E-4
ing segmentation 9.42903E-4
test perplexity 9.415160000000001E-4
class equivalent 9.413000000000001E-4
own class 9.41203E-4
estimation problem 9.38549E-4
corresponding class 9.37395E-4
perplexity values 9.36819E-4
conditional likelihood 9.25157E-4
class dmtnbutmn 9.25018E-4
class ditribu 9.25018E-4
length phrases 9.23219E-4
clustering process 9.05276E-4
parameter estimation 9.05197E-4
perplexity results 8.94E-4
perplexity value 8.894360000000001E-4
ing grammar 8.8818E-4
different pruning 8.84286E-4
probability 8.79323E-4
same clustering 8.7685E-4
viterbi algorithm 8.72811E-4
viterbi perplexity 8.54708E-4
estimation scheme 8.48488E-4
topic identifica 8.39007E-4
other hand 8.34361E-4
perplexity measure 8.30379E-4
training protocol 8.236949999999999E-4
