different feature 0.0030139669999999998
such features 0.002899934
feature set 0.002770066
bow features 0.002717236
feature representations 0.00265635
additional features 0.002636579
amt feature 0.00263175
relevant features 0.002609654
crowdsourced features 0.002578691
deep features 0.002574468
automatic feature 0.002571272
unigram features 0.002568056
bigram features 0.002544014
expert features 0.002536129
crowdsourced feature 0.002535261
feature com 0.002532711
feature sets 0.0025321759999999997
discontinuous features 0.002526907
feature map 0.002524355
feature combina 0.002523959
approximate feature 0.002519684
feature combination 0.002519057
feature weighting 0.002515979
feature combinations 0.002499561
expert feature 0.002492699
features 0.00229904
feature 0.00225561
only word 0.002077414
word classes 0.001934518
word forms 0.001833069
stop words 0.001418891
tinuous words 0.0014180759999999999
other models 0.0014158999999999999
baseline model 0.001389511
model bow 0.0013633809999999999
experiments data 0.001358927
available data 0.00132428
data points 0.001240669
data the 0.0012349799999999999
unigram model 0.001214201
words 0.00118699
same learning 0.00114106
machine learning 0.001030825
review text 0.0010157500000000002
such representations 0.001001634
learning algorithm 0.001000042
other applications 9.83369E-4
gram models 9.80324E-4
answer pairs 9.73794E-4
such problems 9.5007E-4
model 9.45185E-4
answer quality 9.43593E-4
results experiment 9.275049999999999E-4
other participants 9.149379999999999E-4
standard representation 9.11052E-4
other player 9.013679999999999E-4
nlp tasks 9.01213E-4
standard representations 8.87734E-4
ning text 8.812990000000001E-4
bow baseline 8.62522E-4
positive class 8.61011E-4
negative evaluation 8.547699999999999E-4
notation task 8.49544E-4
weighting problem 8.31945E-4
such occurrences 8.30033E-4
small set 8.29994E-4
possible way 8.29559E-4
kernel representations 8.18398E-4
set exp 8.16484E-4
document classification 8.10988E-4
error detection 8.098900000000001E-4
average time 8.03883E-4
new method 8.02873E-4
class dis 8.02465E-4
sentiment analysis 7.97954E-4
closed class 7.93609E-4
class items 7.91709E-4
pos tags 7.91522E-4
error reduction 7.895080000000001E-4
class item 7.8926E-4
maximum number 7.74428E-4
simple baseline 7.7249E-4
standard rep 7.71912E-4
amt figure 7.70721E-4
manual work 7.66071E-4
mean number 7.629E-4
classification experiments 7.57795E-4
classification problems 7.547739999999999E-4
only differences 7.54709E-4
error reduc 7.54284E-4
good representations 7.478000000000001E-4
first attempt 7.445619999999999E-4
important baseline 7.44386E-4
models 7.43425E-4
analysis problems 7.39877E-4
review datasets 7.366639999999999E-4
large margin 7.35545E-4
sentiment annotation 7.3268E-4
bow exp 7.20224E-4
web interface 7.19583E-4
