training data 0.002598861
words data 0.0024327050000000003
data set 0.002286792
different training 0.002025701
data sets 0.0019918130000000003
whole data 0.00192854
irbc data 0.001905343
balanced data 0.00189025
rbc data 0.001885128
imbalanced data 0.001882818
different results 0.001873547
anced data 0.001864739
different methods 0.001821046
different set 0.001713632
training set 0.0016323330000000001
frequency information 0.001508371
different kernels 0.001405441
small training 0.0014022330000000001
function words 0.001400189
different character 0.001397684
language models 0.001363661
different documents 0.001344179
training sets 0.001337354
different authors 0.001332085
different locations 0.001329976
test set 0.0013182760000000002
different loca 0.001298586
different configurations 0.001297238
different parts 0.001292781
different posi 0.001292781
training documents 0.00126288
same classifier 0.001251359
training example 0.001249701
training instances 0.001244035
training examples 0.0012436510000000001
frequent word 0.0012352539999999999
word histograms 0.001230725
training doc 0.001210552
training instance 0.001210552
classification methods 0.001196488
ing words 0.001187624
sequential information 0.0011731089999999999
word usage 0.0011629610000000001
words representations 0.0011560540000000001
common words 0.00115497
same topic 0.001153777
experimental results 0.001145597
style information 0.0011452889999999999
tree methods 0.0011395490000000001
useful information 0.001138239
preliminary information 0.001123224
ter information 0.001118801
quential information 0.0011176229999999999
words characters 0.001110874
other values 0.001108168
show results 0.001102863
positive results 0.001102856
other kernels 0.001092058
kernel function 0.0010824860000000001
parameters words 0.001072961
mental results 0.001059976
representative results 0.001059976
term frequency 0.0010590550000000002
balanced corpus 0.001044995
same pattern 0.001034215
set size 0.001032074
other hand 0.001029386
same locations 0.0010255960000000001
fication methods 0.0010109680000000001
ter methods 0.001010558
weighted frequency 9.97105E-4
same imbalance 9.9517E-4
other researchers 9.8804E-4
kernel distance 9.775270000000002E-4
training 9.72201E-4
language 9.58006E-4
frequency weighting 9.43602E-4
distance measures 9.417340000000001E-4
distribution function 9.40469E-4
test samples 9.370520000000001E-4
similar terms 9.137990000000001E-4
set sizes 9.06631E-4
model selection 9.05859E-4
test docu 9.043510000000001E-4
comparable performance 8.974690000000001E-4
bow approach 8.97222E-4
previous work 8.95665E-4
recognition performance 8.91876E-4
method parameters 8.883560000000001E-4
large number 8.80519E-4
superior performance 8.76143E-4
information 8.75789E-4
linear function 8.7427E-4
such rep 8.61077E-4
smoothing function 8.60558E-4
document representation 8.596739999999999E-4
text summarization 8.56888E-4
similarity function 8.45798E-4
acceptable performance 8.435280000000001E-4
many classes 8.38442E-4
