<?xml version="1.0" standalone="yes"?>
<Paper uid="P06-4010">
  <Title>Chinese Named Entity and Relation Identification System</Title>
  <Section position="6" start_page="39" end_page="39" type="concl">
    <SectionTitle>
5 Conclusion
</SectionTitle>
    <Paragraph position="0"> This three-stage IE prototype system CHINERIS is propriate and entit and relation identifi In the first component, it is a beneficial exploration to develop an error repairer which simultaneously enhances the performance of Chinese rd segmentation and POS tagging.</Paragraph>
    <Paragraph position="1"> In the second component, we theoretically extend the original definition of Finite State Automata (FSA), that is, we use complex mbols rather than atomic constraint symbols.</Paragraph>
    <Paragraph position="2"> With this extension, we improve the practicability for the FSC mechanism. At the same time, the new issue for automatically constructing FSC also increases the flexibility of its maintenance. In order to improve the NE identification performance, some special strategies for the identification of NEs without trigger words are added in this stage, which cannot be recognized by FSC. In the third component, automatically selecting effectual multi-level linguistic features for each NER and non-NER and learning two oppoe types of cases simultaneously are two innovative points in the PNCBL approach.</Paragraph>
    <Paragraph position="3"> The lexical sports ontology plays an important role in the identification of NEs and NERs, such as determination of the boundary of N ation for NE with special constructons and calculation of similarity for the features (e.g. semantic distance).</Paragraph>
    <Paragraph position="4"> The experimental results for the three components in the prototype system show that the system CHINERIS is ication.</Paragraph>
  </Section>
class="xml-element"></Paper>