sentiment classification 0.0036098199999999997
negative sentiment 0.0033172699999999998
sentiment classifier 0.003223333
positive sentiment 0.002940736
sentiment analysis 0.0028986759999999998
general sentiment 0.002807762
semantic sentiment 0.0026907669999999997
sentiment clas 0.0025813959999999997
mixed sentiment 0.002579539
sentiment classifica 0.0025672209999999997
classification training 0.00239948
sentiment 0.0022803
training data 0.00219207
training set 0.002055769
negative text 0.001908589
svm polarity 0.001843459
same topic 0.0017539259999999998
test data 0.0017162649999999998
same set 0.0016692450000000002
timent classification 0.001652125
negative reviews 0.001620467
svm classifier 0.0016194920000000002
topic dependency 0.001606394
positive text 0.001532055
problem text 0.00151224
training testing 0.001492375
coverage polarity 0.0014811260000000001
newswire polarity 0.00148073
other domain 0.001458281
data sets 0.0014530889999999999
noisy data 0.001451056
new set 0.0014455470000000002
same approach 0.001443012
mixed data 0.001421349
newsgroup data 0.001412956
data source 0.0014111619999999999
general text 0.001399081
machine classifier 0.001398879
ent data 0.001391436
training examples 0.001391181
ing features 0.001390179
mix training 0.0013895560000000001
movie review 0.00138317
bayes classifier 0.0013773050000000001
negative distributions 0.0013763739999999999
negative sets 0.0013679489999999998
set size 0.001342275
training sizes 0.001340271
classification 0.00132952
machine learning 0.001322827
next set 0.001301014
feature types 0.001297214
different parameter 0.001265641
analysis problem 0.001258997
unigram features 0.001244053
positive reviews 0.001243933
same classifiers 0.001213323
traditional text 0.001207694
domain dependency 0.001206603
own text 0.001181226
few features 0.001176426
text clas 0.001172715
polarity 0.001167
learning techniques 0.001164275
test documents 0.001164184
movie reviews 0.001152749
other top 0.001111235
topic 0.00107049
training 0.00106996
other extensions 0.001053618
product reviews 0.001011693
svm newswire 9.90189E-4
feature 9.88397E-4
previous work 9.80151E-4
best performance 9.79303E-4
mediocre performance 9.708259999999999E-4
svm fin 9.629230000000001E-4
support vector 9.56419E-4
work this 9.454660000000001E-4
classifier 9.43033E-4
test sets 9.25134E-4
model 9.21909E-4
high level 9.21259E-4
vector machines 9.1513E-4
ing process 9.006940000000001E-4
test domains 9.00229E-4
electronic methods 8.99092E-4
future work 8.79971E-4
optimised results 8.78476E-4
various test 8.75513E-4
emoticons dataset 8.71602E-4
language style 8.697690000000001E-4
learning 8.66981E-4
good match 8.54574E-4
features 8.51185E-4
mean accuracy 8.50904E-4
reserved reviews 8.48294E-4
internet movie 8.381840000000001E-4
ing techniques 8.362879999999999E-4
preliminary experiments 8.339020000000001E-4
