opinion word 0.00517737
opinion target 0.0047630699999999995
opinion words 0.004715668
opinion extraction 0.004653082
chinese opinion 0.004463926
other opinion 0.00440594
opinion analysis 0.004353536
opinion retrieval 0.004350012
same opinion 0.004233435
relevant opinion 0.004230323
opinion relevance 0.004226407
opinion scores 0.004201096
opinion weight 0.004197273
opinion score 0.0041896469999999995
problem opinion 0.004160948
individual opinion 0.004054856
above opinion 0.004053592999999999
opinion prediction 0.004036239
unified opinion 0.004035227
opinion identification 0.0040329269999999995
opinion question 0.004028179
opinion nature 0.00401926
current opinion 0.004015515
opinion summarization 0.004013223
opinion holder 0.004008131
opinion favorite 0.004005655
opinion identifica 0.003990819
opinion generation 0.003990703999999999
complete opinion 0.0039886539999999995
opinion holders 0.003987717
effective opinion 0.003986944
vant opinion 0.0039859629999999995
irrelevant opinion 0.003984688
levant opinion 0.00398435
subsequent opinion 0.003983691
opinion summa 0.003983691
opinion 0.00371677
sentiment word 0.00239822
many word 0.002135969
word lexicon 0.002100081
same word 0.001977265
weight word 0.001941103
sentiment words 0.001936518
opinionated word 0.001877584
word pairs 0.001820624
target topic 0.00178148
word pair 0.001775745
word distance 0.00177545
comment word 0.0017464240000000001
such words 0.001742787
minimum word 0.001737774
topic words 0.001734078
other words 0.001688068
chinese sentiment 0.001684776
sentiment lexicon 0.001577101
relevant target 0.001559853
negative words 0.001559243
information retrieval 0.0015491889999999999
retrieval model 0.0015139440000000001
sentiment corpus 0.0015130719999999999
different approaches 0.001506933
negative sentiment 0.001497965
different documents 0.001494076
different targets 0.001486685
different evaluation 0.0014833329999999999
target topics 0.001482417
corresponding target 0.001440102
evaluation results 0.001416737
target subject 0.0014103639999999999
relevance model 0.001390339
retrieval method 0.00138638
sentiment lexicons 0.001366161
positive words 0.001358363
retrieval performance 0.0013488670000000001
iteration algorithm 0.001342556
need target 0.0013348610000000001
content words 0.0013335719999999999
ent target 0.001330868
contextual information 0.001330261
different context 0.001323092
seed words 0.001320343
identical target 0.001318818
ranking algorithm 0.001317292
ment words 0.0012898010000000001
idea different 0.001288619
comment words 0.001284722
training sentiment 0.001284404
traditional information 0.001282946
hits algorithm 0.0012748310000000001
performance evaluation 0.0012743960000000001
other approaches 0.0012715410000000002
expansion methods 0.001267771
same topic 0.0012518450000000001
topic terms 0.001250218
tive sentiment 0.001248534
different sources 0.001248436
many documents 0.001244883
topic relevance 0.001244817
identification different 0.001240719
good performance 0.001236297
