A88-1016 the following parameters : * The sampling frequency for generated frames
A00-2005 k bootstrap replicates of L by sampling m items from with replacement
A97-1015 domains . We also find the small sampling problem in this experiment .
A00-2040 alternative , based on Monte Carlo sampling of the rules . For each error
A97-1015 In other words , it is a small sampling problem , which can be seen in
C00-1010 derivations are simply removed from the sampling distribution . After sampling
A00-2040 with a ` lazy ' strategy with sampling size 5 ( lazy ( 5 ) + in figure
A00-2009 distribution is created by randomly sampling 349 sense-tagged examples from
A00-2040 for an algorithm based on rule sampling , this effect is much less severe
A00-2040 required . We experimented with sampling sizes 5 and 10 . As CPU requirements
A00-2040 for the ` lazy ' algorithm with sampling size 5 , show that the phoneme
C00-1010 sampling distribution . After sampling a large number of random derivations
A92-1037 process was checked by random sampling , and 92 % of the links was recognized
C00-1010 which can only be enforced by sampling valid representations from the
A00-1030 by Philip Resnik , using some sampling tools that he developed . The
C00-1011 will not use Bod 's Monte Carlo sampling technique from complete derivation
C00-1011 parse both by means of Monte Carlo sampling and by means of Viterbi n-best
A88-1016 dictionary is specific to the sampling frequency used ( 8 , 10 or 16kHz
A00-2041 large space efficiently by first sampling widely and shal - lowly , and
A00-1030 of an explicit determiner . 3.2 Sampling rule performance The morphological
hide detail