<?xml version="1.0" standalone="yes"?>
<Paper uid="W00-1210">
  <Title>A trainable method for extracting Chinese entity names and their relations</Title>
  <Section position="4" start_page="70" end_page="70" type="concl">
    <SectionTitle>
4. Conclusion
</SectionTitle>
    <Paragraph position="0"> In this paper we presented a trainable method for extracting Chinese entity names and their relations. The method provides a unified framework based on MBL. Our preliminary experiment demonstrates that this trainable method is efficient and flexible. Any linguistic features, either surface or deep, can be easily added into the system. Preliminary experiments have shown that our performance is comparable to or better than other existing trainable methods, such as HMM and Crystal. Our work, however, is still in its preliminary stage. More thorough evaluation is required using larger testing corpora. Some algorithmic extensions are also expected so as to improve the performance, including automatic feature selection, coreference resolution, etc.</Paragraph>
  </Section>
class="xml-element"></Paper>