other features 0.002824431
different features 0.002768506
several features 0.002573067
new features 0.002440971
string features 0.0024324439999999998
linguistic features 0.0023659659999999997
syntactic features 0.002353297
feature weight 0.002334685
feature vector 0.0023346269999999997
theoretic features 0.002333273
integration features 0.002332688
individual features 0.00229034
positional features 0.00228428
motivated features 0.002262244
informativeness features 0.002260098
encyclopedic features 0.002257134
new feature 0.002248191
string feature 0.0022396639999999997
binary feature 0.002129751
gain feature 0.002115076
feature con 0.002114373
novel feature 0.002114245
numeric feature 0.002100603
feature importance 0.0020744839999999997
features 0.00202353
feature 0.00183075
training data 0.0017126250000000002
data set 0.001559956
test data 0.001535691
training set 0.001512485
learning algorithm 0.0014858979999999998
different learning 0.001455426
text sentence 0.001436355
learning method 0.00141338
word set 0.001373975
other documents 0.0013699
other work 0.001346551
same data 0.001339336
classification algorithm 0.001325998
same training 0.0012918650000000001
other phrases 0.001271554
ing data 0.001261558
other baseline 0.001249615
word frequency 0.001237452
many discourse 0.001228606
large number 0.0012117249999999999
training candidate 0.001202127
original text 0.00119941
data sets 0.001194203
other methods 0.00119359
press data 0.0011901400000000001
input text 0.001178844
large documents 0.0011594560000000001
ing algorithm 0.001156958
anchor text 0.001155476
learning process 0.00114597
understanding text 0.001142758
put text 0.001142758
word extraction 0.001140276
traditional information 0.001138569
learning algorithms 0.001135494
training collection 0.001131259
supervised algorithm 0.0011258499999999999
data sparsity 0.0011193050000000001
other systems 0.001107669
learning methods 0.001103139
standard information 0.001102863
ratio training 0.00110038
other count 0.001091397
mutual information 0.00109117
first evaluation 0.001079823
information gain 0.001078253
keyword set 0.001077009
document frequency 0.001067311
ing set 0.001061418
supervised learning 0.001060852
discourse properties 0.001053574
supervised method 0.0010533320000000001
classification decision 0.0010497969999999999
discourse comprehension 0.00104597
comprehension discourse 0.00104597
information retrieval 0.001045107
important words 0.0010413570000000001
bayes learning 0.001040183
phrases documents 0.001039652
dic information 0.0010279220000000001
mit information 0.0010279220000000001
common word 0.001025338
extraction system 0.001022954
learning scheme 0.001014509
term frequency 0.001011901
machine learning 0.001009076
genetic algorithm 0.001008916
national corpus 0.001002971
test instances 0.001000413
word counts 9.97998E-4
classifier performance 9.95303E-4
phrases document 9.94579E-4
high score 9.92772E-4
discourse compre 9.90485E-4
