#30416This paper explores the issue of using different co-occurrence similarities betweenterms for separating query terms that are useful for retrieval from those that are harmful.
tech,20-1-E99-1034,ak
query terms
</term>
that are useful for
<term>
retrieval
</term>
from those that are harmful . The
#30425This paper explores the issue of using different co-occurrence similarities between terms for separating query terms that are useful forretrieval from those that are harmful.
measure(ment),3-3-E99-1034,ak
</term>
. Preliminary experiments with
<term>
similarities
</term>
computed using
<term>
first-order and
#30457Preliminary experiments withsimilarities computed using first-order and second-order co-occurrence seem to confirm the hypothesis.
other,6-3-E99-1034,ak
<term>
similarities
</term>
computed using
<term>
first-order and second-order co-occurrence
</term>
seem to confirm the hypothesis .
<term>
#30460Preliminary experiments with similarities computed usingfirst-order and second-order co-occurrence seem to confirm the hypothesis.
other,9-4-E99-1034,ak
then be used for determining which
<term>
query terms
</term>
are useful and best reflect the
<term>
#30479Term similarities could then be used for determining whichquery terms are useful and best reflect the user's information need.
other,17-4-E99-1034,ak
</term>
are useful and best reflect the
<term>
user 's information need
</term>
. A possible application would be
#30487Term similarities could then be used for determining which query terms are useful and best reflect theuser 's information need.
measure(ment),14-5-E99-1034,ak
source of evidence for tuning the
<term>
weights
</term>
of the
<term>
query terms
</term>
. We
#30506A possible application would be to use this source of evidence for tuning theweights of the query terms.