<?xml version="1.0" standalone="yes"?>
<Paper uid="A00-2030">
  <Title>A Novel Use of Statistical Parsing to Extract Information from Text</Title>
  <Section position="11" start_page="231" end_page="232" type="evalu">
    <SectionTitle>
10 Experimental Results
</SectionTitle>
    <Paragraph position="0"> Our system for MUC-7 consisted of the sentential model described in this paper, coupled with a simple probability model for cross-sentence merging. The evaluation results are summarized in Table 1.</Paragraph>
    <Paragraph position="1"> In both Template Entity (TE) and Template Relation (TR), our system finished in second place among all entrants. Nearly all of the work was done by the sentential model; disabling the cross-sentence model entirely reduced our overall F-Score by only 2 points.  While our focus throughout the project was on TE and TR, we became curious about how well the model did at part-of-speech tagging, syntactic parsing, and at name finding. We evaluated part-of-speech tagging and parsing accuracy on the Wall Street Journal using a now standard procedure (see Collins 97), and evaluated name finding accuracy on the MUC-7 named entity test. The results are summarized in Table 2.</Paragraph>
    <Paragraph position="2"> While performance did not quite match the best previously reported results for any of these three tasks, we were pleased to observe that the scores were at or near state-of-the-art levels for all cases.</Paragraph>
  </Section>
class="xml-element"></Paper>