unlabeled data 0.002496181
training data 0.0024477780000000003
word frequency 0.002192677
classification data 0.00218061
word tokens 0.002076748
labeled data 0.002051181
data set 0.002047245
other data 0.002040741
marginal word 0.002022837
word counts 0.00196957
word representa 0.001949532
word occurrences 0.001947313
word marginals 0.001930315
mass word 0.001920644
word freq 0.001915103
data sets 0.001903172
reuters data 0.0018756880000000001
entire data 0.001812886
separate data 0.0017892160000000002
mod data 0.0017890360000000001
particular data 0.001786255
data stores 0.0017682230000000002
unlabeled corpus 0.0015506370000000001
many words 0.001489998
learning algorithm 0.001476421
unlabeled examples 0.001434792
training size 0.001410426
unlabeled corpora 0.0013976280000000002
labeled training 0.0013962990000000002
training set 0.001392363
training examples 0.001386389
learning methods 0.0013279099999999999
small training 0.001322244
unique words 0.00131766
learning performance 0.001314388
negative training 0.001289025
unknown words 0.001287095
mon words 0.001271211
training sets 0.00124829
unlabeled dataset 0.001242291
supervised learning 0.001203487
unlabeled documents 0.001189357
feature value 0.001186326
sparse training 0.001185354
many features 0.00117464
positive class 0.001167608
large corpus 0.001166669
feature frequency 0.001165497
feature set 0.0011637549999999998
bayes model 0.0011190940000000002
machine learning 0.001112956
labeled corpus 0.001105637
class probability 0.001102458
text classification 0.001088328
frequency information 0.001082519
classification task 0.001068434
positive test 0.001065933
original feature 0.001065136
large set 0.0010567979999999999
words 0.0010528
negative class 0.00103388
test examples 0.0010295690000000001
novel learning 0.001017955
unigram features 0.001014062
large corpora 0.00101366
target class 0.001007658
feature space 0.001003659
manifold learning 0.001001898
classification tasks 9.894069999999999E-4
sentiment classification 9.80983E-4
detailed results 9.76557E-4
classification accuracy 9.728709999999999E-4
sparse feature 9.56746E-4
primary results 9.48458E-4
experimental results 9.41317E-4
ing set 9.401489999999999E-4
classification threshold 9.39462E-4
different weighting 9.29818E-4
novel algorithm 9.265020000000001E-4
ranking results 9.22942E-4
small set 9.21711E-4
topic classification 9.12628E-4
document classification 9.09043E-4
classification setting 9.06961E-4
feature marginals 9.03135E-4
training 8.96448E-4
different meth 8.93659E-4
feature spaces 8.93432E-4
new task 8.916309999999999E-4
class mcat 8.867759999999999E-4
ing task 8.833879999999999E-4
negative classes 8.833370000000001E-4
negative examples 8.825180000000001E-4
tive class 8.779409999999999E-4
details class 8.75516E-4
ative class 8.71866E-4
standard label 8.71412E-4
model 8.65107E-4
earnings class 8.64263E-4
same way 8.601399999999999E-4
