different training 0.001913708
data set 0.0017236309999999998
training set 0.001583438
possible attachment 0.001563231
noun attachment 0.0014805670000000001
high attachment 0.001457691
attachment statistics 0.001423482
attachment disambiguation 0.001421718
attachment errors 0.001408547
attachment ambiguity 0.001403288
attachment decision 0.001402726
low attachment 0.0013850210000000002
same set 0.001379996
test set 0.001365139
corpus size 0.001363357
clause attachment 0.0013453850000000002
data sets 0.001340368
evaluation results 0.0013247929999999999
default attachment 0.001321864
attachment phrase 0.001319541
data structure 0.001318779
frequent attachment 0.001316277
different sub 0.001309259
different lattices 0.0013073919999999999
likely attachment 0.001303357
attachment lbd 0.001303327
attachment node 0.001298923
attachment train 0.001285114
attachment decisions 0.00128347
large training 0.001279412
phrase data 0.001277012
attachment points 0.001276351
reuters corpus 0.001274067
unlabeled corpus 0.0012656590000000001
attachment nodes 0.001259899
attachment ambiguities 0.0012538900000000001
attachment possibilities 0.001250372
attachment ambigui 0.001250372
data problems 0.001247575
annotated corpus 0.0012443110000000001
small training 0.001242999
train data 0.001242585
unannotated corpus 0.0012406840000000001
wsj corpus 0.001229204
unannotated data 0.0012275519999999998
labeled corpus 0.001226761
disambiguation information 0.001226147
sparse data 0.001211364
efficient data 0.0012079839999999998
other work 0.001202864
training sets 0.001200175
mutual information 0.001179316
many sentences 0.001134227
set size 0.001118034
disambiguation system 0.0011047869999999999
other dependencies 0.001100106
maximum training 0.001099746
statistical parser 0.001081589
ing set 0.001077265
wsj training 0.001075879
pervised training 0.00107395
dependency statistics 0.001071476
partial dependency 0.001061068
lbd results 0.0010286240000000001
important dependency 0.0010281140000000001
lexicalized parser 0.001024646
disambiguation method 0.0010231349999999999
methods collins 0.001022756
new text 0.001021997
attachment 0.00102044
system subj 0.001018881
recent results 0.001016434
development set 9.98198E-4
set sizes 9.97878E-4
parameter set 9.948119999999999E-4
corpus 9.91043E-4
experimental results 9.89611E-4
collins parser 9.85532E-4
pendency parser 9.83191E-4
other difference 9.82134E-4
test sets 9.81876E-4
dependency structures 9.80432E-4
poor results 9.79534E-4
ambiguous sentences 9.78792E-4
query results 9.77621E-4
disambiguation performance 9.729249999999999E-4
monetary system 9.68434E-4
tistical parser 9.67009E-4
combined system 9.65247E-4
baseline method 9.622529999999999E-4
other combina 9.51276E-4
ambiguous context 9.36543E-4
bined system 9.345639999999999E-4
relation mod 9.342269999999999E-4
good performance 9.27661E-4
much context 9.13786E-4
lexical statistics 8.98767E-4
limited context 8.98434E-4
syntactic disambiguation 8.96647E-4
large number 8.84475E-4
