<?xml version="1.0" standalone="yes"?>
<Paper uid="W03-0424">
  <Title>Language Independent NER using a Maximum Entropy Tagger</Title>
  <Section position="7" start_page="0" end_page="0" type="concl">
    <SectionTitle>
6 Conclusion
</SectionTitle>
    <Paragraph position="0"> Our NER system demonstrates that using a large variety of features produces good performance. These features can be defined and extracted in a language independent manner, as our results for German, Dutch and English show. Maximum entropy models are an effective way of incorporating diverse and overlapping features. Our maximum entropy tagger employs Gaussian smoothing which allows a large number of sparse, but informative, features to be used without overfitting.</Paragraph>
    <Paragraph position="1"> Using a wider context window than 2 words may improve performance; a reranking phase using global features may also improve performance (Collins, 2002).</Paragraph>
  </Section>
class="xml-element"></Paper>