previous word 0.0030592270000000003
word classes 0.0030453930000000004
word vocabulary 0.0029423060000000004
current word 0.0029228860000000004
preceding word 0.0028977020000000003
unseen word 0.002897494
word com 0.00289024
word triggers 0.0028899610000000004
word sequences 0.002889635
language model 0.002887399
word combinations 0.0028594730000000004
known word 0.002848629
likely word 0.002846458
word combina 0.002842546
word combi 0.0028425050000000004
meaningful word 0.002841792
word sample 0.002841225
sand word 0.002841225
bigram model 0.002393221
trigram model 0.002382678
markov model 0.002373856
language models 0.002363059
backoff model 0.002286503
baseline model 0.002254056
order model 0.002218402
unigram model 0.002210906
variable model 0.0022048750000000002
gram model 0.002192992
combined model 0.0021764790000000003
aggregate model 0.002168405
entropy model 0.002167945
base model 0.002166592
model 0.00191912
bigram models 0.001868881
trigram models 0.001858338
such models 0.00185724
markov models 0.001849516
different words 0.001834815
backoff models 0.001762163
many words 0.0017123239999999999
typical models 0.001696547
order models 0.001694062
unigram models 0.001686566
variable models 0.001680535
gram models 0.001668652
aggregate models 0.0016440650000000001
previous words 0.0016407869999999999
models scales 0.001622275
bust models 0.001619488
common words 0.001505726
frequent words 0.001472931
statistical language 0.001449922
training data 0.001434439
models 0.00139478
language modeling 0.001372838
training corpus 0.0013069800000000001
test data 0.001302472
training sentence 0.001301085
natural language 0.001279543
language mod 0.0012477249999999999
training set 0.001226978
words 0.00119757
test perplexity 0.001164452
probability distribution 0.001156434
sentence probabilities 0.001125383
set perplexity 0.0011150069999999999
probability figure 0.001106432
large number 0.0010991109999999998
test set 0.001095011
assignment probability 0.001080605
single class 0.001076525
overall training 0.0010683350000000001
probability estimates 0.001066768
class label 0.001063205
training time 0.001059146
different values 0.001051475
same sentence 0.00105004
interpolation algorithm 0.0010492330000000001
probable class 0.001044345
class assignments 0.0010287360000000001
long training 0.001021301
different dependencies 0.0010206730000000002
probability mass 0.001003007
posterior probability 9.96897E-4
probability esti 9.96824E-4
finite probability 9.94621E-4
leftover probability 9.92951E-4
language 9.68279E-4
sparse data 9.6512E-4
different approaches 9.62418E-4
overall perplexity 9.56364E-4
training times 9.49973E-4
ing set 9.418059999999999E-4
training efficiency 9.413010000000001E-4
backoff test 9.39611E-4
different order 9.36527E-4
arge number 9.3078E-4
feasible training 9.290430000000001E-4
different learning 9.264900000000001E-4
total number 9.198139999999999E-4
