<?xml version="1.0" standalone="yes"?>
<Paper uid="W00-1417">
  <Title>OLAP context Dimensions</Title>
  <Section position="5" start_page="130" end_page="130" type="concl">
    <SectionTitle>
4 Conclusion
</SectionTitle>
    <Paragraph position="0"> We presented a new approach to content aggregation in the context of a very challenging and practical generation application: summarizing OLAP and data mining discoveries 13t as a few linked web pages of fluent and concise natural language. We believe that the key contribution to our work is to show the feasibility to perform effective paratactic aggregation: (r) encapsulated within a single generation component (the sentence planner) @ simple data .structure,'- me -~- raetonzauon ........... index.html matrix, that captures the key structural and ordering constraints on paratactic aggregation while completely abstracting from domain semantic idiosyncrasies as well as from lexicai and syntactic details.</Paragraph>
    <Paragraph position="1"> This is a first success towards the development of a plug-in content aggregation component for text generation, reusable across application domains. In future work, we intend to empirically evaluate the summaries generated by HYSSOP.</Paragraph>
  </Section>
class="xml-element"></Paper>