word vector 0.00337308
word vectors 0.002840291
negative sentiment 0.002516777
sentiment classification 0.002496586
sentiment information 0.002434913
positive sentiment 0.0023203959999999997
twitter sentiment 0.002254567
sentiment labels 0.002233719
sentiment analysis 0.002216452
target vector 0.002189081
model training 0.002187003
correct sentiment 0.002186371
vector representations 0.00213767
adaptive sentiment 0.002136987
sentiment propagations 0.002095148
sentiment classifica 0.002092317
dependent sentiment 0.002088146
feature vector 0.001971967
vector representation 0.00187427
sentiment 0.00185254
binary vector 0.001819999
bias vector 0.0017327900000000001
truth vector 0.001716232
words features 0.0016595520000000001
negative words 0.0016471440000000001
semantic composition 0.001524872
composition function 0.00146384
recursive neural 0.00143903
training data 0.0014344660000000001
vector 0.00143011
different composition 0.00141975
neural models 0.001352248
target information 0.001341344
model 0.00133583
composition matrix 0.0012995300000000001
learning models 0.0012984759999999998
child vectors 0.001262734
semantic categories 0.001257074
opinion words 0.001234531
connected words 0.0012220170000000002
target node 0.001214717
softmax function 0.001186904
different parameter 0.001177309
dependency tree 0.0011678679999999999
root vectors 0.001159231
semantic compositions 0.001154007
syntactic relationships 0.001151556
syntactic tags 0.00114388
movie review 0.001138713
classification dataset 0.00113776
objective function 0.001136331
combined vectors 0.001134647
semantic compo 0.001124629
training split 0.001094654
recursive structure 0.001093239
training instance 0.001082441
powerful function 0.001077337
nonlinearity function 0.001072565
composition functions 0.001068717
binary features 0.001066534
linearity function 0.001064694
neural network 0.0010598
polarity classification 0.001054029
method accuracy 0.001045222
target expansion 0.001025827
different targets 0.001024511
negative softmax 0.001017852
adaptive recursive 0.001016993
composition process 0.0010146
multiple composition 0.001009258
composition selection 9.93338E-4
interested target 9.90229E-4
specified target 9.88601E-4
twitter data 9.8532E-4
words 9.82907E-4
neural mod 9.78046E-4
dependency results 9.68063E-4
tree input 9.64421E-4
build recursive 9.62423E-4
composition matrices 9.58375E-4
stanford parser 9.574290000000001E-4
selection method 9.55291E-4
global matrix 9.4938E-4
content features 9.36269E-4
negative sentiments 9.34851E-4
binary feature 9.31746E-4
several entities 9.30326E-4
review dataset 9.228540000000001E-4
adaptive composition 9.14998E-4
dependent features 9.12251E-4
converted tree 9.087119999999999E-4
ment classification 9.08598E-4
pendency tree 9.06123E-4
tree output 9.028739999999999E-4
subjectivity classification 9.028E-4
dency tree 9.00545E-4
deep learning 8.9993E-4
other settings 8.97676E-4
vectors 8.97321E-4
dependency parsing 8.80571E-4
