A00-2035 |
and POS tags . This is why the "
|
lexical lookup
|
" method is the major source
|
A92-1004 |
subclass fields are filled via
|
lexical lookup
|
. The lexicon is organized into
|
A00-2035 |
is implemented by applying the
|
lexical lookup
|
method where a word is assigned
|
A00-2035 |
the error rate from 15 % of the
|
lexical lookup
|
approach to 5 % . This suggests
|
A88-1013 |
new lexical en - tries . In the
|
lexical lookup
|
stage , an acceptability threshold
|
A88-1013 |
performed on the base form , and
|
lexical lookup
|
is performed after each transformation
|
C92-1014 |
save disk accesses by limiting
|
lexical lookups
|
of intermediate forms to just
|
C92-1040 |
A CCG deriw ~ tion starts with
|
lexical lookup
|
, which identifies the functional
|
C86-1035 |
attack ) phase . 2 . We assume that
|
lexical lookup
|
for the entire sentence is done
|
C00-1057 |
standardized form . This facilitates
|
lexical lookup
|
and allows tile system to map
|
C92-1014 |
inflectional analyzer when to attempt a
|
lexical lookup
|
of an intermediate form stem
|
A00-2035 |
which is an improvement over the
|
lexical lookup
|
approach by 2-3 times . On the
|
A88-1013 |
of candidate fixes found during
|
lexical lookup
|
was not ordered or pruned before
|
C88-1058 |
string , the first steps being
|
lexical lookup
|
and counting and indexing all
|
C00-1008 |
slight modification of sl ; andard
|
lexical lookup
|
, lexical entries are al ) propriately
|
C94-2127 |
step , whic : h corresponds to
|
lexical lookup
|
in chart parsers , is to look
|
E03-2014 |
system performs tokenisation ,
|
lexical lookup
|
, and HMM-POS disambiguation
|
C00-1057 |
improved . Cohunn A provides baseline
|
lexical lookup
|
and parsing times based on untagged
|
A94-1045 |
of a system which must perform
|
lexical lookup
|
on a vast dictionary , followed
|
A88-1013 |
, and can be implemented as a
|
lexical lookup
|
routine which heuristically prunes
|