sentiment classification 0.003079
polarity classification 0.00284099
classification model 0.0024801809999999997
text classification 0.002399498
training set 0.002390506
training model 0.0023173309999999997
sentiment classifier 0.002315007
training data 0.0022825140000000002
classification accuracy 0.002156666
polarity words 0.001997413
classification systems 0.001984802
classification models 0.00196712
classification algo 0.001931289
ment classification 0.001910445
classification sys 0.00185253
dual training 0.001817544
training sample 0.001811936
learning methods 0.0018060160000000001
sentiment orientation 0.001804698
training samples 0.0018046569999999999
opposite training 0.0017980539999999999
data set 0.00179306
sentiment datasets 0.001759919
sentiment classifica 0.001750136
sentiment classifi 0.001735042
feature vector 0.0016826929999999999
classification 0.00160283
polarity shift 0.001575949
machine learning 0.001552333
handling polarity 0.001532988
polarity classi 0.001519801
polarity reversion 0.00151131
polarity shifters 0.0015083190000000002
dling polarity 0.0014881540000000002
sentiment 0.00147617
learning systems 0.001472812
word order 0.001459317
test set 0.001457244
training 0.00143998
learning algorithms 0.001430865
new word 0.0014255539999999999
learning errors 0.001413875
negation word 0.001407917
linguistic features 0.001371826
ing data 0.001360039
test data 0.001349252
positive text 0.001327314
feature engineering 0.001313371
itive feature 0.001307126
orientations word 0.0012909290000000001
original data 0.001288537
representation model 0.0012873119999999999
features dataset 0.001281881
polarity 0.00123816
ing method 0.001220059
text representation 0.001206629
unigram features 0.001184562
effective features 0.0011743259999999998
bigrams features 0.0011730579999999998
negative indicator 0.00116958
bow model 0.0011618639999999999
class label 0.001161628
such approach 0.001148574
bigram features 0.0011463509999999999
semantic information 0.001144491
dataset svm 0.001138301
negation words 0.00112788
learning 0.00109084
different systems 0.001089704
following results 0.001067219
ment words 0.001066868
shift problem 0.001066796
popular text 0.001060845
feature 0.00105233
whole text 0.001051748
svm naïve 0.0010467200000000001
alternative methods 0.0010157579999999999
following rules 0.00100362
tive methods 9.98775E-4
such dis 9.91649E-4
experimental results 9.76415E-4
different domains 9.658589999999999E-4
other com 9.580139999999999E-4
related work 9.55707E-4
original test 9.527209999999999E-4
positive score 9.39373E-4
current work 9.18001E-4
other alterna 9.16657E-4
product reviews 9.146689999999999E-4
subsequent work 8.84007E-4
features 8.81358E-4
test sample 8.78674E-4
model 8.77351E-4
baseline system 8.71456E-4
test samples 8.71395E-4
opposite test 8.64792E-4
strong pos 8.630129999999999E-4
bayes baseline 8.58974E-4
linear kernel 8.568479999999999E-4
classifier 8.38837E-4
