<?xml version="1.0" standalone="yes"?>
<Paper uid="P95-1054">
  <Title>Quantifying lexical influence: Giving direction to context</Title>
  <Section position="8" start_page="332" end_page="333" type="concl">
    <SectionTitle>
5 Summary and Future Work
</SectionTitle>
    <Paragraph position="0"> The preliminary results described in this work establish clearly that non-standard metrics of lexical  direction have been selected randomly from the test-set. Note that very few of these pairs exhibit comparable influence on each other. The arrows indicate the direction of lexical influence (or information flow). A DIM score of 1 or more implies a significant association, whereas an MIS below 4 is considered a chance association. influence bear much promise. In fact, what we really need is a generalised information score, a measure that takes into account several factors, such as:  The generalised information score would capture all the variations that are introduced by the above factors, and allow for the variants so as to reflect a &amp;quot;normalised&amp;quot; measure of contextual influence.</Paragraph>
    <Paragraph position="1"> We have also been working with experimental measures which attach higher significance to the collocation frequency, (measures which, in essence, &amp;quot;trust&amp;quot; the recogniser more often). Our future work will involve bringing these various factors together into one integrated formalism.</Paragraph>
  </Section>
class="xml-element"></Paper>