prototype word 0.0028679549999999997
word type 0.002831753
word types 0.002714738
quent word 0.0026026699999999996
biguate word 0.0025933019999999996
unsupervised model 0.002411694
frequent words 0.001775548
model 0.00173823
rare words 0.001596301
context vectors 0.001589285
clustering algorithm 0.0015292629999999999
training data 0.00138134
context matrix 0.001369959
words 0.0013558
other models 0.001303006
gold tag 0.001300412
unsupervised pos 0.001285153
hmm models 0.001260774
context distributions 0.001232089
unsupervised learning 0.001230229
cluster number 0.0011774139999999999
different tags 0.001175514
context matrices 0.001170023
deterministic algorithm 0.001169656
right context 0.001164039
data set 0.001159878
same type 0.001155209
markov models 0.001123977
original context 0.0011009000000000001
latent features 0.001090433
left context 0.001089931
current models 0.001083943
available models 0.001082405
ntypes context 0.001079513
hmm training 0.001076807
unsupervised algorithms 0.001064675
prototype tags 0.001046007
final tag 0.0010342489999999999
same label 0.001028513
entire corpus 0.001028419
unsupervised way 0.00101633
annotated data 0.001009268
pos tagging 0.0010063749999999999
important feature 0.0010033870000000001
annotated corpus 9.923830000000001E-4
constituent vectors 9.76047E-4
clustering algorithms 9.750449999999999E-4
scriptor vectors 9.7213E-4
clustering step 9.60845E-4
training times 9.607099999999999E-4
results section 9.56966E-4
algorithm 9.45429E-4
common gold 9.32267E-4
cluster centroids 9.276999999999999E-4
single matrix 9.112650000000001E-4
unsupervised taggers 9.08688E-4
guistic features 9.0267E-4
final clustering 9.01455E-4
unsupervised gram 9.01414E-4
large number 9.00082E-4
learning tasks 8.893729999999999E-4
other algorithms 8.785609999999999E-4
tagging accuracy 8.767289999999999E-4
accuracy scores 8.56657E-4
context 8.40081E-4
clustering algo 8.17728E-4
models 8.15656E-4
first iteration 8.10277E-4
gold stan 8.074029999999999E-4
first pass 8.038559999999999E-4
descriptor matrix 8.02703E-4
ferent tokens 7.890449999999999E-4
hmm state 7.85883E-4
difficult learning 7.799339999999999E-4
svd method 7.72605E-4
various types 7.49848E-4
vectors 7.49204E-4
new pair 7.46397E-4
recent methods 7.418380000000001E-4
main obstacle 7.33156E-4
corpus 7.32766E-4
bigram counts 7.32511E-4
large values 7.310089999999999E-4
future work 7.29856E-4
singular value 7.23086E-4
important characteristics 7.20982E-4
present work 7.16427E-4
original work 7.09641E-4
ging accuracy 7.083549999999999E-4
several hours 7.05566E-4
dot product 7.0107E-4
value decomposition 6.994869999999999E-4
tagging problem 6.975379999999999E-4
normalization step 6.955290000000001E-4
standard deviations 6.95071E-4
cluster 6.93153E-4
full wall 6.91937E-4
weighted average 6.91049E-4
original space 6.89093E-4
latent descriptor 6.88824E-4
