<?xml version="1.0" standalone="yes"?>
<Paper uid="X98-1018">
  <Title>DYNAMIC DATA FUSION</Title>
  <Section position="3" start_page="0" end_page="0" type="relat">
    <SectionTitle>
PREVIOUS WORK
</SectionTitle>
    <Paragraph position="0"> The ability to improve retrieval performance by using multiple retrieval systems has been documented extensively (e.g,, \[1\], \[31, \[41). It is only recently, however, that researchers have turned their attention to the possibility of adjusting the manner in which results are combined to the specific query at hand. Researchers have reported success in using initial relevance judgments to adjust the way in which results are combined \[3\], and have also reported success in using the joint distribution of relevance scores from multiple marchers (among other things) to predict when to combine the results of multiple systems \[6\]. The purpose of the current research is to explore the use of the joint distribution of relevance scores, semantic and syntactic features of queries.</Paragraph>
    <Paragraph position="1"> and the length of retrieved documents to predict how to combine the results of several retrieval systems.</Paragraph>
  </Section>
class="xml-element"></Paper>