conditional model 0.0022624950000000002
entropy model 0.0022258070000000002
inequality model 0.002129961
joint model 0.002092057
above model 0.0020876080000000003
model expectations 0.002067073
model the 0.002044614
equality model 0.002021735
tional model 0.002017171
feature function 0.002012289
uniform model 0.002010923
extended model 0.0020066290000000002
model avoids 0.001999949
many features 0.0018736920000000002
model 0.00176233
active features 0.00166746
possible features 0.00165345
feature expectations 0.0016492730000000001
feature selection 0.001649046
rare features 0.001599006
sparse features 0.0015835200000000002
tive features 0.001582709
unnecessary features 0.0015616220000000001
liable features 0.001559709
training data 0.001460471
feature 0.00134453
features 0.00132156
following models 0.001301646
inequality models 0.0012686519999999999
gaussian models 0.001254092
equality models 0.001160426
same time 0.001138028
ing method 0.001137141
objective function 0.001106763
other nlp 0.0010707450000000001
other methods 0.001048082
gaussian distribution 0.001045652
many optimization 0.001043972
training examples 0.001039981
such tasks 0.0010387
optimization problem 0.001027446
probability space 0.001005894
entropy estimation 9.9405E-4
bayes method 9.90069E-4
language modeling 9.89016E-4
many parameters 9.75984E-4
lagrange method 9.74853E-4
lmvm method 9.5686E-4
map estimation 9.55467E-4
original probability 9.484739999999999E-4
posterior probability 9.447629999999999E-4
uniform distribution 9.411739999999999E-4
conditional maximum 9.394760000000001E-4
other criteria 9.33036E-4
beta distribution 9.319179999999999E-4
terior distribution 9.319179999999999E-4
tive function 9.28908E-4
many nlp 9.28598E-4
maximum entropy 9.027880000000001E-4
models 9.01021E-4
maximum number 8.96408E-4
following optimization 8.924650000000001E-4
standard maximum 8.87767E-4
empirical results 8.79672E-4
several tasks 8.77844E-4
bayesian estimation 8.66136E-4
inequality constraint 8.64967E-4
constraint width 8.631380000000001E-4
evaluation set 8.63E-4
many times 8.55494E-4
optimization problems 8.5431E-4
ment set 8.48075E-4
dual problem 8.477210000000001E-4
several algorithms 8.473250000000001E-4
efficient estimation 8.46875E-4
single value 8.46523E-4
development set 8.385809999999999E-4
different widths 8.34397E-4
estimation algorithms 8.33082E-4
several situations 8.240300000000001E-4
small number 8.23074E-4
blmvm algorithm 8.15557E-4
iis algorithm 8.12419E-4
gis algorithm 8.12419E-4
nlp datasets 8.08362E-4
text categorization 8.015310000000001E-4
variable metric 8.000559999999999E-4
constraint widths 7.953560000000001E-4
optimization algorithms 7.94349E-4
several solutions 7.8647E-4
above results 7.842960000000001E-4
estimation tool 7.825810000000001E-4
bust estimation 7.780339999999999E-4
gaussian map 7.77965E-4
first half 7.760040000000001E-4
empirical expectation 7.74374E-4
width factor 7.64107E-4
natural combination 7.64081E-4
slack variables 7.62509E-4
inequality constraints 7.59125E-4
