opinion entity 0.004059831
opinion extraction 0.003909283
opinion expression 0.003768415
opinion analysis 0.003599521
opinion expressions 0.003592518
standard opinion 0.003558033
opinion lexicon 0.003546905
opinion recognition 0.003531965
opinion attribute 0.003485359
extracts opinion 0.003446073
opinion expres 0.003443724
automatic opinion 0.003442801
giant opinion 0.003433015
opinion 0.0031703
extraction features 0.001980543
common features 0.0015863490000000001
transition features 0.001565178
mon features 0.001505981
polarity extraction 0.001482406
feature vectors 0.001466459
feature functions 0.001439208
negative method 0.00137534
learning approach 0.0013584550000000002
expression extraction 0.001337098
entity level 0.001336528
features 0.00124156
low method 0.001192293
experimental results 0.001171205
polarity annotations 0.001162202
tagging method 0.001160423
feature 0.00115471
learning models 0.001140697
polarity component 0.001131565
different test 0.001128368
evaluation results 0.0011089049999999999
attribute method 0.001101724
joint extraction 0.00110093
polarity labels 0.001100218
sequential learning 0.001089664
subjective expression 0.001087918
previous work 0.001085997
correct polarity 0.001078998
separate sequence 0.001077138
intensity extraction 0.001076562
only model 0.0010722729999999999
method description 0.001070939
polarity attributes 0.001069558
test dataset 0.001068455
polarity transition 0.001067041
tagging approach 0.0010637210000000001
empirical results 0.0010624929999999999
polarity attribute 0.001058482
much data 0.001056694
ging method 0.00105237
joint approach 0.00105191
polarity none 0.001037072
sequence taggers 0.00102752
strength extraction 0.001023508
boundary extraction 0.001004087
such dis 0.001004057
separate set 9.92368E-4
current sentence 9.91427E-4
same compo 9.89107E-4
nent approach 9.58099E-4
negative classes 9.502969999999999E-4
hierarchical parameter 9.498600000000001E-4
ferent set 9.39678E-4
quential learning 9.3907E-4
individualized learning 9.33005E-4
negative class 9.274E-4
negative intensity 9.26254E-4
ion expression 9.1288E-4
parameter sharing 9.060299999999999E-4
sentence cor 9.034360000000001E-4
hierarchical structure 9.00304E-4
related work 8.95886E-4
other classes 8.904449999999999E-4
individual performance 8.895740000000001E-4
entity 8.89531E-4
conditional random 8.87933E-4
same subcomponents 8.69524E-4
joint task 8.58768E-4
recognition task 8.584860000000001E-4
other variations 8.55799E-4
ing scheme 8.4488E-4
ous work 8.346580000000001E-4
work studies 8.3321E-4
test instance 8.29925E-4
conditional probability 8.294190000000001E-4
hierarchical sequential 8.27024E-4
vious work 8.17092E-4
parameter shar 8.16923E-4
performance degradation 8.09509E-4
attribute classification 8.06759E-4
recognition systems 7.99427E-4
weight vectors 7.950559999999999E-4
sequential tagging 7.9493E-4
entire test 7.93338E-4
end system 7.92182E-4
key terms 7.87886E-4
