lexical features 0.003305867
new features 0.0031185350000000004
many features 0.0030788900000000004
dialog features 0.002940078
significant features 0.0029346550000000004
various features 0.002913793
structural features 0.002842429
boolean features 0.002839397
lex features 0.002837665
thrnew features 0.00283085
verbosity features 0.002828829
noisy features 0.002828829
positional features 0.002828829
imt features 0.002828829
features 0.00261404
feature set 0.0024831949999999997
feature class 0.002409726
new feature 0.002347005
feature sets 0.00213903
indicator feature 0.0021102359999999997
feature versions 0.002083147
feature name 0.002075959
structural feature 0.0020708989999999997
feature subsets 0.0020691999999999998
boolean feature 0.0020678669999999997
feature denot 0.002057704
ious feature 0.002057704
feature 0.00184251
such data 0.0015791339999999998
multiple data 0.0012704259999999998
system performance 0.001263236
data point 0.001246832
learning system 0.001215207
data points 0.001205839
data statistics 0.001198639
test set 0.001189663
other person 0.0011878219999999998
different levels 0.001180904
different subsets 0.00116541
different slices 0.001155636
new set 0.00114518
word unigrams 0.001093373
word ngram 0.001092343
test time 0.001090913
word ngrams 0.001080464
lexical patterns 9.928580000000001E-4
similar results 9.90148E-4
such cases 9.82083E-4
other forms 9.697429999999999E-4
pos tag 9.626890000000001E-4
social power 9.50454E-4
other people 9.35698E-4
new problem 9.32256E-4
evaluation set 9.29433E-4
automatic system 9.28501E-4
lexical cues 9.13551E-4
social context 9.057609999999999E-4
supervised learning 8.95461E-4
class baseline 8.860840000000001E-4
learning sys 8.789850000000001E-4
dev set 8.75859E-4
first message 8.69715E-4
train set 8.69167E-4
erence set 8.67362E-4
high accuracy 8.617449999999999E-4
such threads 8.47906E-4
test sets 8.45498E-4
accuracy baseline 8.428209999999999E-4
majority class 8.42806E-4
future work 8.316860000000001E-4
email corpus 8.30421E-4
basic nlp 8.232809999999999E-4
participant pairs 8.192659999999999E-4
lex system 8.04432E-4
best system 7.97094E-4
second person 7.95813E-4
manual annotation 7.93134E-4
power relations 7.848130000000001E-4
dev test 7.84152E-4
words 7.7794E-4
power relation 7.76647E-4
average number 7.76482E-4
pos tags 7.725119999999999E-4
prediction task 7.68872E-4
tion task 7.68517E-4
blind test 7.66154E-4
email messages 7.58547E-4
many emails 7.57546E-4
description accuracy 7.56868E-4
cal power 7.48412E-4
enron email 7.482560000000001E-4
email thread 7.461620000000001E-4
previous studies 7.44466E-4
text samples 7.4424E-4
organizational email 7.35063E-4
thread level 7.19302E-4
message count 7.18082E-4
social networks 7.18062E-4
ing participant 7.13986E-4
email cor 7.10543E-4
