word translation 0.00251027
labeled data 0.001832171
training data 0.0017980919999999998
word pair 0.001767123
word segmentation 0.001740314
model accuracy 0.001733487
word pairs 0.001711785
simple word 0.00170808
learning method 0.0016947379999999999
word correlations 0.0016925479999999999
japanese word 0.001660092
word trans 0.001659773
word correspon 0.001655726
query word 0.001654867
language sentiment 0.0016525060000000002
source data 0.001649002
new feature 0.001601091
learning task 0.0015977219999999998
unlabeled data 0.001580265
source domain 0.001571762
source language 0.0015652920000000002
model estimation 0.0015622359999999998
feature space 0.0015587399999999999
feature vector 0.00155687
model complexity 0.001550182
target domain 0.00154637
target language 0.0015399
machine translation 0.001538775
translation baseline 0.001501166
domain adaptation 0.001489758
adaptation domain 0.001489758
language text 0.001473865
sentiment classification 0.0014362799999999998
data upper 0.001425756
labeled training 0.001421903
structural learning 0.0014144799999999999
classification performance 0.001407124
other features 0.001404004
unsupervised domain 0.001402933
translation oracle 0.001394836
dimensional feature 0.001383592
natural language 0.0013766520000000001
language use 0.0013742900000000002
each feature 0.001372239
single translation 0.0013716
feature barrier 0.0013592959999999999
classification task 0.001354226
correspondence learning 0.001346968
classification problem 0.0013425759999999998
language processing 0.001340774
translation technology 0.001340267
learning rate 0.001336285
discriminative learning 0.0013349829999999999
chine translation 0.001326241
translation noise 0.001320479
translation technol 0.001320479
language correspondence 0.0013196980000000002
basic domain 0.001316702
classification results 0.001315108
learning setup 0.001307228
standard classification 0.00130093
classification problems 0.001294205
model 0.00129105
learning paradigm 0.0012826579999999999
spondence learning 0.0012813299999999998
sentiment analysis 0.0012736129999999998
ferent language 0.001273414
separate domain 0.0012669410000000002
domain adap 0.0012613680000000001
domain adapta 0.001260511
text classification 0.001257639
label words 0.0012537569999999999
classification tasks 0.001241891
training set 0.0012028709999999999
other words 0.00120056
such problems 0.001193967
different source 0.001168603
different approach 0.00116434
different classifier 0.001156427
different target 0.0011432109999999999
training linear 0.00114034
other methods 0.001138649
training error 0.0011352419999999998
labeled documents 0.001131471
classification knowledge 0.0011247129999999998
feature 0.00111955
performance results 0.0011137439999999998
training documents 0.001097392
training reviews 0.0010909819999999999
translation 0.00108811
different languages 0.0010845569999999999
classification studies 0.001083716
semantic analysis 0.001081996
parameter matrix 0.001080381
target task 0.001069412
classification prob 0.001057767
ment classification 0.0010546829999999998
learning 0.00104774
empirical analysis 0.001043572
classification example 0.001042106
