data feature 0.003723007
feature selection 0.003529473
feature set 0.003510391
different feature 0.003493925
same feature 0.003382165
feature weights 0.00330709
feature scores 0.003293399
feature space 0.003287119
feature combinations 0.00324227
good feature 0.003227215
candidate feature 0.003188788
feature expectation 0.003176982
final feature 0.003176968
feature count 0.003174693
incremental feature 0.003164518
various feature 0.003159599
feature groups 0.0031456459999999998
cme feature 0.003143972
expected feature 0.003140156
combined feature 0.003137565
feature list 0.0031360199999999998
feature variables 0.003126955
splitting feature 0.003122537
progressive feature 0.003120233
feature subspaces 0.003118809
feature spaces 0.003114332
unlimited feature 0.003112994
feature selec 0.003107427
empty feature 0.003106153
feature sub 0.00310501
huge feature 0.003104925
predefined feature 0.003102886
feature 0.00289311
new features 0.002593438
good features 0.0023787549999999998
additional features 0.002344425
tional features 0.0023145699999999997
features precision 0.0022881259999999997
copy features 0.002279241
prosodic features 0.002259088
ble features 0.0022551679999999997
features 0.00204465
training data 0.001612573
selection algorithm 0.0015998779999999999
other words 0.0015667559999999999
new algorithm 0.001512303
current model 0.0013871839999999999
conditional model 0.0013815639999999998
entropy model 0.001320614
same data 0.001318952
exponential model 0.001318052
test data 0.0012910719999999999
cme model 0.0012653019999999998
data split 0.001241585
model construction 0.0012277079999999999
lected model 0.001226015
pfs algorithm 0.001220653
sgc algorithm 0.001215767
cme algorithm 0.001214377
tion algorithm 0.001213572
computation algorithm 0.001200108
words distance 0.001196176
available algorithm 0.001188117
partial word 0.001178437
ifs algorithm 0.00117714
word flags 0.001176732
word rough 0.00116176
switchboard data 0.001147833
syntactic information 0.0011476540000000001
word fragment 0.00114434
word fragments 0.00114434
wnr word 0.001140977
same set 0.0011063359999999999
training process 0.001092551
new corpus 0.001068549
velopment data 0.001060306
data subsets 0.001055685
training algo 0.0010548180000000001
positions words 0.0010490550000000001
orthographic words 0.001047715
segmented data 0.001040126
ened words 0.0010383130000000001
filler words 0.0010383130000000001
new models 0.001033876
selection algorithms 0.00103254
contextual information 0.001028506
training instance 0.001017128
other machine 0.001015162
model 0.00101444
different split 0.001012503
progressive training 0.001009799
same accuracy 0.0010067639999999998
other variations 0.00100446
fast training 9.93888E-4
prosody label 9.84415E-4
other variables 9.738699999999999E-4
algorithm 9.63515E-4
same gain 9.61649E-4
other alternatives 9.52382E-4
selection process 9.462380000000001E-4
