translation model 0.00378269
function word 0.0033978970000000004
word alignments 0.0030930920000000004
word order 0.003046136
word frequency 0.003006501
word penalty 0.002945389
word relationships 0.0029070570000000002
chinese word 0.002897858
word list 0.0028603
nal word 0.0028504460000000004
terrogative word 0.0028079430000000002
word fre 0.002807309
structed word 0.002806734
model feature 0.002755376
language model 0.002697057
english translation 0.002585145
dominance model 0.0024629409999999997
ing model 0.002415703
nance model 0.002271086
gram model 0.002264376
translation performance 0.002250037
model our 0.002249873
lexical translation 0.0022453910000000002
inance model 0.002240011
translation probability 0.002218025
machine translation 0.002118221
correct translation 0.0021156910000000003
translation tasks 0.00209922
translation errors 0.0020428490000000002
translation quality 0.002036706
incorrect translation 0.00203616
model 0.00200995
jective translation 0.002003075
hierarchical phrase 0.001996229
function words 0.001904777
phrase features 0.001903157
phrase alignments 0.001888512
source sentence 0.00179896
noun phrase 0.001781328
translation 0.00177274
phrase penalty 0.0017408089999999998
target language 0.001673517
source language 0.001663373
phrase marker 0.00165993
uence phrase 0.001602715
target phrases 0.001596495
training data 0.0015246539999999999
tion words 0.00143347
alignment information 0.001390601
source side 0.001383086
system models 0.001374054
phrase 0.00137222
arabic source 0.001370091
frequent words 0.0013671170000000002
consistent alignment 0.001353794
source text 0.001333227
enough words 0.001329447
tent words 0.0013294180000000002
cal alignment 0.001321774
other function 0.001315415
hierarchical rules 0.001305789
same set 0.0012836570000000001
target languages 0.001276957
alignment points 0.001266567
alignment heuris 0.001258758
english corpus 0.001251941
hierarchical phrases 0.001234094
second function 0.001224435
english side 0.001219225
previous function 0.001217122
different approach 0.001214905
same rule 0.001208385
dominance feature 0.001198417
sentence pairs 0.001178694
other rules 0.001176098
syntactic information 0.001147042
chinese sentence 0.001143752
syntactic con 0.001139681
test set 0.001132925
correct reordering 0.0011133839999999998
derivation models 0.001112703
approximate function 0.001109376
hierarchical structure 0.001107595
other phrases 0.001104403
syntactic categories 0.001102626
syntactic modeling 0.0010965929999999999
language side 0.001093927
linguistic phrases 0.001091356
ned function 0.00108796
words 0.00108368
neighboring function 0.001073442
motivated function 0.001070738
individual function 0.001064715
reordering choice 0.001060447
human language 0.001051432
data sparseness 0.001034492
incorrect reordering 0.001033853
reordering choices 0.0010313359999999999
full training 0.001028896
syntactic roles 0.001022522
