phrase length 0.002767995
phrase extraction 0.002689027
caller phrase 0.002573937
phrase labels 0.002443302
phrase counts 0.0024398320000000003
phrase lengths 0.0023790540000000002
actual phrase 0.00235709
phrase 0.00210302
word error 0.0020981000000000003
word indices 0.002018293
training data 0.002003488
word hypoth 0.0019986780000000003
discrete word 0.0019986780000000003
caller phrases 0.001660087
candidate phrases 0.001652508
lexical features 0.0016355419999999998
many phrases 0.001618832
evaluation data 0.001586759
many features 0.0015783919999999999
hzp model 0.001570546
development data 0.001549151
training corpus 0.001516077
information extraction 0.001495176
trigram features 0.001435731
event model 0.001429402
phrases most 0.001417353
cal features 0.0014068849999999999
gram features 0.001404914
lexical information 0.0013959810000000001
own data 0.0013958199999999999
opment data 0.001380608
data sparseness 0.001380587
caller information 0.001380086
sitional features 0.0013791819999999998
intrinsic features 0.001375519
other words 0.001327446
small set 0.001261951
automatic speech 0.001257379
other number 0.001218193
model 0.00119614
phrases 0.00118917
ing training 0.001181346
different tasks 0.001172957
relevant information 0.001164248
features 0.00114873
sitional information 0.001139621
performance results 0.0011388
main results 0.0011381109999999998
sic information 0.001136118
development set 0.001128866
number extraction 0.001127609
extraction task 0.001124299
speech recognizer 0.001121338
entity extraction 0.001089347
speech recognizers 0.001084023
different techniques 0.001080985
length distribution 0.00108087
extraction grammar 0.001078721
small number 0.001072708
development corpus 0.00106174
corpus development 0.00106174
feature inventory 0.001060892
different corpora 0.0010550590000000001
other methods 0.001044896
voicemail corpus 0.00103159
different event 0.001031169
phone number 0.001027667
main approach 0.001023797
other sources 0.00100856
candidate number 0.00100494
simple extraction 0.001000127
common words 9.97235E-4
same message 9.87729E-4
classifier method 9.845259999999999E-4
name length 9.81233E-4
tree classifier 9.78678E-4
such basic 9.68446E-4
possible extraction 9.6374E-4
accurate set 9.622000000000001E-4
rich set 9.604440000000001E-4
baseline approach 9.57732E-4
extraction methods 9.54312E-4
candidate phone 9.494029999999999E-4
good performance 9.3984E-4
other settings 9.37818E-4
other means 9.373840000000001E-4
large number 9.371869999999999E-4
number domain 9.32268E-4
such taggers 9.27312E-4
hzp models 9.19802E-4
other hand 9.105750000000001E-4
information 9.09169E-4
other kinds 9.06579E-4
name extraction 9.02265E-4
same representative 9.021680000000001E-4
classification task 9.004060000000001E-4
prietary corpus 9.00103E-4
language processing 8.87777E-4
good news 8.85942E-4
first names 8.84681E-4
