training data 0.0022781
linear feature 0.001915391
feature functions 0.001899136
training set 0.001867437
feature product 0.001864738
feature func 0.001862005
feature expectations 0.001848953
quadratic feature 0.00184172
feature covariances 0.001820649
unlabeled data 0.0018018920000000003
crf training 0.0017925430000000002
training approach 0.001769324
unlabeled training 0.001747752
training algorithm 0.001743448
large data 0.0017365420000000002
crf model 0.001728663
labeled data 0.001674474
sequence data 0.0016546010000000001
labeled training 0.0016203340000000002
feature 0.00159987
new training 0.001599316
data sets 0.0015953830000000001
training time 0.00159209
training method 0.001521088
supervised training 0.001506048
conditional model 0.00150374
training methods 0.001460589
output data 0.001454863
data sequences 0.0014449430000000002
supervised model 0.001442168
testing data 0.0014247410000000002
training algo 0.001421791
discriminative training 0.001415927
whole data 0.0014139500000000002
training procedure 0.001408056
training objective 0.001405832
data points 0.0013922740000000002
unlabeled set 0.001391229
nearby data 0.0013904610000000002
possible label 0.0013748850000000002
training criterion 0.001374498
efficient training 0.0013609100000000001
training process 0.001358823
available training 0.001354807
initial model 0.001344904
effective training 0.001341035
training procedures 0.001340808
training technique 0.0013380830000000001
crf approach 0.0013379070000000002
self training 0.001337715
model parameters 0.001337144
training iterations 0.001333259
combined training 0.001331126
training proce 0.001331126
arbitrary features 0.0013311240000000002
rare features 0.0013279230000000002
correlated features 0.0013268910000000002
crf algorithm 0.001312031
erful model 0.0012684859999999999
labeled set 0.001263811
label sequence 0.0012396030000000001
structured label 0.001206736
other words 0.0011508389999999999
new approach 0.00114468
structured crf 0.001136177
unlabeled sequence 0.001124253
new algorithm 0.0011188040000000002
crf models 0.001115844
training 0.00111198
features 0.00110558
previous label 0.001100391
conditional probability 0.001081603
test sequence 0.001079183
supervised crf 0.001074631
label classification 0.001073578
test time 0.001070812
unlabeled sets 0.001065035
standard crf 0.001050448
model 0.0010481
word sense 0.001046302
corresponding label 0.001041719
learning approach 0.0010300510000000001
label sequences 0.001029945
supervised algorithm 0.001025536
class label 0.001025315
same accuracy 0.0010112950000000002
word boundaries 0.001011013
ing sets 0.001008702
learning algorithm 0.001004175
labeled sequence 9.968350000000002E-4
optimization algorithm 9.9482E-4
crf figure 9.90652E-4
regularization function 9.812409999999999E-4
finite label 9.794180000000001E-4
beled set 9.77328E-4
seed set 9.75863E-4
bined set 9.74954E-4
crf objective 9.74415E-4
label alphabet 9.71886E-4
large corpus 9.64496E-4
