C04-1127 |
pattern candidates ordered by the
|
ranking function
|
. The resulting performance is
|
C04-1202 |
the explicit form of sentence
|
ranking functions
|
for the automatic text summarization
|
C04-1199 |
reimplementation does not use a
|
ranking function
|
( Radev et al. , 2000 ) . When
|
D08-1018 |
We discuss the details of the
|
ranking function
|
f used to compute the utility
|
D08-1015 |
retrieval . These techniques generate
|
ranking functions
|
which predict the relevance of
|
D08-1015 |
. First , all of the pairwise
|
ranking functions
|
are applied to the unseen document
|
D08-1018 |
Section 5.3 . Using utility as the
|
ranking function
|
, we sort all pairs of symbols
|
C04-1202 |
learning mechanism for sentence
|
ranking function
|
. In the preliminary experiments
|
D08-1018 |
be written as O ( Nkn3 ) . 5.3
|
Ranking function
|
We discuss the details of the
|
D08-1015 |
the M ( M -- 1 ) / 2 pairwise
|
ranking functions
|
gjk created during the training
|
C04-1202 |
standard deviation . 2.3 Sentence
|
ranking function
|
We assume that for a certain
|
D08-1018 |
determine the better form of the
|
ranking function
|
f as well as to tune its weights
|
C04-1202 |
for an ideal uniform sentence
|
ranking function
|
. 6 Acknowledgements Our thanks
|
D09-1068 |
most important feature sets in
|
ranking functions
|
. Traditional proximity features
|
D09-1055 |
associations into a search engine 's
|
ranking function
|
should help improve both search
|
C04-1143 |
length constraints . The following
|
ranking function
|
rank ( j ) , where j is the sentence
|
D09-1068 |
approach to web search is to learn a
|
ranking function
|
h ( xi ) , where xi is a feature
|
C04-1202 |
is scored based on a sentence
|
ranking function
|
constructed by GEP . Fitness
|
D08-1111 |
is w * , then we can make the
|
ranking function
|
as f ( x ) = ( w * , x ) . When
|
C04-1202 |
the current candidate sentence
|
ranking function
|
under consideration : ) ) , (
|