A00-3002 |
ignoring their length . The second
|
weighting
|
scheme we tried was based on
|
A92-1022 |
longer hy - potheses . Statistical
|
weighting
|
by itself produced an improvement
|
A00-1019 |
the sentence . A study of the
|
weightings
|
for these two models is described
|
A00-2011 |
account . This is achievable by
|
weighting
|
the translation scores by the
|
A88-1007 |
would detract from the overall
|
weighting
|
. This would allow the system
|
A00-1040 |
according to formula 1 where is a
|
weighting
|
constant often set to 1 . A full
|
A88-1007 |
observed patterns as one factor in
|
weighting
|
. Frequently occurring patterns
|
A92-1025 |
The following is the formula for
|
weighting
|
terms that are individual indicators
|
A00-1022 |
decision rules or probability
|
weightings
|
. During the categorization phase
|
A00-3002 |
found by the parser . The second
|
weighting
|
scheme is more fine-grained and
|
A00-2038 |
Heeringa ( 1997 ) experimented with
|
weighting
|
each feature by information gain
|
A92-1026 |
experimentation with var ions
|
weightings
|
, we concluded that the length
|
A00-3002 |
extraction time for the first
|
weighting
|
scheme was below 5 % of the total
|
A00-3002 |
7.5 = 14.5 . Although the first
|
weighting
|
scheme seems rather crude , it
|
A00-1038 |
and General , is used to make a
|
weighting
|
distinction between these and
|
A00-1038 |
terms , contextual information ,
|
weighting
|
, if-then rules and other pattern
|
A88-1033 |
direction will include refining the
|
weighting
|
criteria , extending the method
|
A92-1025 |
techniques we explored include various
|
weighting
|
schemes , with the end goal being
|
A92-1022 |
increased when combined with a
|
weighting
|
scheme which prefers longer hy
|
A00-1025 |
weighting for documents and / tu term
|
weighting
|
for the question ( Singhal et
|