word pairs 0.003485
word pair 0.0030779509999999998
important word 0.0029791569999999996
top word 0.002977464
content word 0.002952924
aggregated word 0.002922735
useful word 0.002922198
gregate word 0.0029028229999999997
other features 0.002750932
lexical features 0.002686382
similar features 0.002490828
binary features 0.002488613
parse features 0.002460192
additional features 0.0024579470000000003
context features 0.002452297
pair features 0.002426241
wordnet features 0.002351719
polarity features 0.002324818
tional features 0.002300107
score features 0.002290766
negation features 0.002289321
sparse features 0.002265484
feature set 0.002227022
features 0.00201898
feature category 0.002008954
feature space 0.001970233
marker feature 0.001969499
single feature 0.001932094
negation feature 0.001896431
feature cat 0.001864897
feature 0.00162609
training data 0.001524103
language model 0.00151591
discourse relations 0.0015154
implicit relation 0.001421762
negative words 0.001339321
content words 0.001338504
functional words 0.001334811
relation types 0.001323688
other work 0.001322885
pdtb relation 0.001313446
relation prediction 0.001306564
hierarchical relation 0.001298027
member words 0.001289439
relation classes 0.001277075
particular relation 0.0012753320000000001
implicit relations 0.0012649129999999999
potential relation 0.001242976
ticular relation 0.001227409
explicit relations 0.001221231
data sets 0.001181963
pdtb data 0.00117617
argument pairs 0.001172059
similar approach 0.001167962
final system 0.0011315700000000001
annotated relations 0.0011194249999999998
data instance 0.001107869
bayes model 0.001107537
large corpus 0.001104901
implicit discourse 0.001103435
other applications 0.001095809
plicit relations 0.001090583
preceding relations 0.001079786
possible approach 0.001072438
entrel relations 0.001071172
words 0.00105627
full system 0.00103779
previous work 0.001034561
best system 0.00103173
same class 0.001029922
discourse marker 0.00102037
other classes 0.001013739
pair set 0.0010081930000000001
syntactic parse 0.001004446
recent work 0.001002228
relation 9.95288E-4
discourse markers 9.91896E-4
semantic relationships 9.89345E-4
work this 9.87929E-4
external corpus 9.85909E-4
content similarity 9.77849E-4
unannotated corpus 9.75934E-4
gigaword corpus 9.75451E-4
contextual information 9.73921E-4
wsj corpus 9.72447E-4
discourse treebank 9.703310000000001E-4
other theories 9.65626E-4
discourse structure 9.65144E-4
our approach 9.546909999999999E-4
approach our 9.546909999999999E-4
aggregated approach 9.48159E-4
information gain 9.42086E-4
powerful approach 9.329449999999999E-4
cosine similarity 9.32308E-4
penn discourse 9.30067E-4
related work 9.246809999999999E-4
discourse rela 9.22994E-4
lexical cues 9.06174E-4
same unsuper 8.75846E-4
future work 8.73915E-4
