other word 0.005039633999999999
several word 0.004930123
word similarity 0.004915717
first word 0.004905441
word frequency 0.004790824
second word 0.004775933
word sense 0.004760802999999999
word pairs 0.004747169999999999
general word 0.004734866
word classes 0.004705633
unseen word 0.00469287
word pair 0.004665071
particular word 0.004651175
olor word 0.004634238
take word 0.004617812999999999
usual word 0.0046145959999999995
ambiguous word 0.0046145959999999995
word cooccurrences 0.0046145959999999995
language model 0.00340227
model error 0.0027769220000000002
similarity model 0.002720927
model performance 0.002525826
discounted model 0.002434377
model 0.00217816
other words 0.002135094
different words 0.002020701
language models 0.002011593
similar words 0.00189237
cluster words 0.001865188
many words 0.0018273410000000001
bigram language 0.001818638
lar words 0.0017494440000000002
lion words 0.001710675
probability estimates 0.001608343
language process 0.001565891
language modeling 0.001562978
unigram probability 0.001545402
statistical language 0.001539377
confusion probability 0.001535168
same probability 0.001522009
probability estimate 0.001519853
base language 0.001515317
natural language 0.001513122
language processing 0.001494626
conditional probability 0.001494171
language mod 0.001487627
probability mass 0.0014804150000000001
words 0.00146841
compact language 0.001465544
training corpus 0.001457969
probability distri 0.001428983
sion probability 0.0014284200000000001
probability ratios 0.00142684
true probability 0.00142684
such models 0.001424831
test corpus 0.0014006420000000001
training data 0.001376106
test data 0.001318779
language 0.00122411
probability 0.00118453
large training 0.001180014
other methods 0.001144619
such methods 0.0011152830000000001
training set 0.001107177
mle models 0.001051219
test set 0.00104985
similarity function 0.0010489879999999998
several features 0.0010474059999999999
other alternative 0.001047358
such events 0.001007242
previous work 9.805740000000001E-4
training sets 9.681749999999999E-4
similarity estimates 9.6658E-4
performance test 9.663899999999999E-4
ing similarity 9.592699999999999E-4
smoothing method 9.57526E-4
error rate 9.52223E-4
similarity work 9.51183E-4
redistribution data 9.508240000000001E-4
other difficulties 9.498849999999999E-4
data sparseness 9.47771E-4
sparse data 9.46523E-4
sentence fragment 9.409189999999999E-4
corresponding training 9.32393E-4
unseen bigram 9.14448E-4
estimation methods 9.13368E-4
test sets 9.108479999999999E-4
other variations 9.106489999999999E-4
pair probabilities 8.88309E-4
base bigram 8.857349999999999E-4
smoothing methods 8.81965E-4
error rates 8.68321E-4
disambiguation method 8.61751E-4
test instance 8.60936E-4
bigram fre 8.5351E-4
appealing approach 8.50246E-4
discounting approach 8.50246E-4
ing set 8.476289999999999E-4
first element 8.427549999999999E-4
conditional distribution 8.41428E-4
