<?xml version="1.0" standalone="yes"?>
<Paper uid="E95-1003">
  <Title>Criteria for Measuring Term Recognition</Title>
  <Section position="4" start_page="19" end_page="19" type="concl">
    <SectionTitle>
4 Conclusion
</SectionTitle>
    <Paragraph position="0"> The term-recognition criteria proposed above - measuring recall and precision for the exact bracketing of maximal termforms- provide a basic minimum of information needed to assess system performance. For some applications, it is useful to further specify how these performance ratios differ for the recognition of simple and complex termforms, how they vary for terms resulting from different term-formation processes, what the ratios are for termform types as opposed to tokens, or how well the system recognizes novel termforms not already in a system lexicon or previously encountered in a training corpus. Precision measurements might usefully state to what extent errors are due to syntactic noise (bracketing crossing syntactic constituents) as distinguished from terminological noise (bracketing including nonclassificatory modifiers or omitting classificatory ones). Publishing such performance results for term-recognition systems would not only display their strengths but also expose their weaknesses. Doing so would ultimately benefit researchers, developers and users of term-recognltion systems.</Paragraph>
  </Section>
class="xml-element"></Paper>