feature learning 0.002881754
input features 0.002486876
textual features 0.0024867089999999997
features cqa 0.0024793799999999998
basic features 0.002466139
representation feature 0.002432433
various features 0.002420056
additional features 0.00241744
textual feature 0.0023734589999999996
excellent features 0.002370816
feature collection 0.00229213
feature sparsity 0.0022698299999999996
features 0.00216784
feature 0.00205459
learning model 0.002023784
quality answer 0.0015503399999999999
answer quality 0.0015503399999999999
regression model 0.001517038
rbm model 0.001472512
graphical model 0.001440597
entropy model 0.001437318
model architecture 0.001426063
shallow model 0.001399348
word frequency 0.001366028
low word 0.001354931
word segmentation 0.001312714
cqa answer 0.001295798
answer length 0.001283297
deep learning 0.00128241
training data 0.0012798380000000002
answer integer 0.001259706
learning approach 0.001254145
frequent words 0.001251956
learning methods 0.001232864
unique words 0.001227582
answer prediction 0.001223854
float answer 0.001216466
model 0.00119662
answer ratio 0.001193668
machine learning 0.001179399
learning framework 0.001152099
learning rate 0.001135855
learning network 0.001126384
learning field 0.001092131
quality answers 0.001086533
same question 0.001084747
learning supervised 0.0010701320000000001
learning architecture 0.001056607
text classification 0.0010404630000000002
different methods 9.98332E-4
classification task 9.89609E-4
words 9.86624E-4
answer 9.84258E-4
classification results 9.75292E-4
textual information 9.75248E-4
data vectors 9.61897E-4
linear classifier 9.593889999999999E-4
several answers 9.54764E-4
high quality 9.50106E-4
training method 9.48413E-4
quality evaluation 9.458279999999999E-4
other methods 9.41678E-4
classification performance 9.39439E-4
question integer 9.328050000000001E-4
testing data 9.265710000000001E-4
data modality 9.16962E-4
other layer 9.06025E-4
question answering 8.83083E-4
new layer 8.64265E-4
question title 8.58691E-4
large number 8.398139999999999E-4
different kinds 8.37925E-4
answers figure 8.3451E-4
low quality 8.303029999999999E-4
training stage 8.298579999999999E-4
learning 8.27164E-4
swer quality 8.18325E-4
training details 8.12585E-4
supervised training 8.10409E-4
hidden vector 8.0914E-4
quality prediction 8.05678E-4
different num 8.020929999999999E-4
deep research 8.00493E-4
following methods 7.840049999999999E-4
other approaches 7.72865E-4
first experiment 7.601280000000001E-4
deep network 7.544660000000001E-4
hidden layer 7.498279999999999E-4
input vector 7.48395E-4
textual vector 7.48228E-4
conditional distribution 7.477040000000001E-4
experimental results 7.38961E-4
integer number 7.36532E-4
average results 7.324289999999999E-4
semantic relevance 7.27737E-4
deep belief 7.2389E-4
complicated answers 7.22658E-4
semantic relationship 7.07004E-4
related work 7.00896E-4
new perspec 6.96415E-4
