<?xml version="1.0" standalone="yes"?>
<Paper uid="W02-0311">
  <Title>Utilizing Text Mining Results: The PastaWeb System</Title>
  <Section position="6" start_page="0" end_page="0" type="concl">
    <SectionTitle>
5 Conclusions and Future Work
</SectionTitle>
    <Paragraph position="0"> We have described PASTAWeb, a prototype Web-based interface to the IE results of the PASTA system. Scientists in the area of molecular biology can benefit from the novel navigation and information tracking capabilities of PASTAWeb and use it as a tool for fast access to specialised information.</Paragraph>
    <Paragraph position="1"> At the time of writing, the database of processed texts accessible through PASTAWeb is rather small (975 texts in total). The rate at which new articles appear on MEDLINE and the limited resources devoted to PASTA make it prohibitive to develop and maintain for PASTAWeb a text database of size comparable to MEDLINE. Nevertheless, PASTAWeb offers the core technology for the development of a fully automated IE system whose input can be based on automated updates (&amp;quot;autoalerts&amp;quot;) from MEDLINE without human intervention. Current work concentrates on the development of such an automated software component and on the feasibility of expanding the system's navigation capabilities to allow users to link together information provided by PASTAWeb and by related servers such as the Protein Data Bank or SWISSPROT.</Paragraph>
    <Paragraph position="2"> Finally, the utility of an interface such as PASTAWeb can only be truly assessed by user evaluation. Usability evaluation should be carried out using both qualitative and quantitative methods. Qualitative evaluation should be used to review the users' perceptions about the design, their preferred strategies for accessing information and whether they find the system easy to use and useful for performing their tasks. Quantitative evaluation should focus on measures of activity time, efficiency in tracking relevant information and on analysing the effect &amp;quot;noise&amp;quot; in the IE results has on user satisfaction. However, while this evaluation remains to be done, we believe that the work presented here provides concrete, constructive ideas about how to effectively utilise the output of IE systems in the biology domain.</Paragraph>
  </Section>
class="xml-element"></Paper>