bigram model 0.0026034110000000004
unigram model 0.0025185190000000003
hierarchical model 0.0025146680000000003
word frequency 0.002302762
word vocabulary 0.002076929
word type 0.002054945
model 0.0020091
word frequencies 0.00198967
word tokenswi 0.001982154
word typew 0.001982154
bayesian models 0.0018594269999999999
such models 0.00175174
hierarchical models 0.0017103779999999998
previous words 0.001692136
base distribution 0.001660636
possible words 0.001575785
form distribution 0.001559335
hdp models 0.0015434
related models 0.001498649
guage models 0.001478596
chical models 0.001460649
language modeling 0.001395851
natural language 0.001372182
inference method 0.0012542030000000002
distribution 0.0012404
base probability 0.00122806
new table 0.001205976
models 0.00120481
table counts 0.001170126
different numbers 0.001140448
predictive probability 0.001135346
explicit table 0.001116934
different frequencies 0.0011072389999999999
words 0.00109451
gibbs sampling 0.001086365
such sampling 0.001073652
table count 0.0010635409999999999
exact table 0.001054859
empirical table 0.001054501
data structure 0.001047834
count algorithm 0.001043074
language 0.00103968
approximate table 0.001033868
sampling methods 0.001027133
sampling process 0.001020529
actual table 0.001015164
efficient method 0.001014541
expected table 0.001008583
bayesian statistics 0.001008007
table assignments 0.001007345
table labels 9.96147E-4
table occupancy 9.90571E-4
old table 9.90552E-4
new customer 9.8962E-4
table tracking 9.863530000000001E-4
approximating table 9.84384E-4
alternative method 9.828E-4
efficient data 9.75929E-4
unsupervised learning 9.60044E-4
token count 9.55118E-4
naive method 9.527170000000001E-4
previous customers 9.44666E-4
bayesian lan 9.35216E-4
hdp bigram 9.329010000000001E-4
base figure 9.323809999999999E-4
many tables 9.23773E-4
several methods 9.211969999999999E-4
hierarchical dirichlet 9.09237E-4
dirichlet process 8.974759999999999E-4
possible seating 8.95662E-4
many customers 8.869800000000001E-4
various counts 8.85293E-4
wsj corpus 8.75979E-4
base distributions 8.74176E-4
standard errors 8.65935E-4
exact values 8.62406E-4
chinese restaurant 8.57914E-4
crp state 8.52599E-4
customer count 8.47185E-4
particular number 8.46575E-4
see figure 8.46411E-4
important information 8.4488E-4
collapsed gibbs 8.33257E-4
total number 8.225729999999999E-4
gibbs samplers 8.21125E-4
values com 8.10475E-4
probability 8.07824E-4
concentration parameter 8.00917E-4
many situations 7.965860000000001E-4
ith customer 7.829E-4
infinite number 7.7763E-4
unigram crp 7.73968E-4
hierarchical pitman 7.65827E-4
procedure increment 7.64913E-4
mcmc sampler 7.645969999999999E-4
constant base 7.64155E-4
based distributions 7.610150000000001E-4
naive approach 7.59211E-4
digamma function 7.55707E-4
yor process 7.495409999999999E-4
