<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-1110">
  <Title>Automated Alignment and Extraction of Bilingual Domain Ontology for Medical Domain Web Search</Title>
  <Section position="4" start_page="0" end_page="0" type="metho">
    <SectionTitle>
3 Evaluation
</SectionTitle>
    <Paragraph position="0"> To evaluate the proposed approach, a medical web search system was constructed. The web pages were collected from several Websites and totally 2322 web pages for medical domain and 8133 web pages for contrastive domain were collected.</Paragraph>
    <Paragraph position="1"> On the other hand, the training and test queries for training and evaluating the system performance were also collected. Forty users, who do not take part in the system development, were asked to provide a set of queries given the collected web pages. After post-processing, the duplicate queries and the queries out of the medical domain are removed. Finally, 3207 test queries mixed Chinese with English words using natural language were obtained.</Paragraph>
    <Section position="1" start_page="0" end_page="0" type="sub_section">
      <SectionTitle>
3.1 Keyword-Based VSM Approach: A
</SectionTitle>
      <Paragraph position="0"> baseline system for comparison In recent years, most of the information retrieval approaches were based on the Vector-Space Model (VSM). Assuming that the query is denoted as a  where a1= . This approach for key term expansion based on synonym set is also adopted in the baseline system. The results and discussions are described in the following sections.</Paragraph>
    </Section>
    <Section position="2" start_page="0" end_page="0" type="sub_section">
      <SectionTitle>
3.2 Weight determination using 11-avgP score
</SectionTitle>
      <Paragraph position="0"> The medical domain web search system is modeled by the linear combination of relational inference model and axiom inference model. The normalized weight factor, a , is employed for concept expansion as follows.</Paragraph>
      <Paragraph position="1">  This experiment is conducted on the estimation of the combination weights for each model. The results are shown in Figure 5. The performance measure called 11-AvgP [Eichmann and Srinivasan 1998] was used to summarize the precision and recall rates. The best 11-AvgP score will be obtained when the weight 0.428a = .</Paragraph>
    </Section>
    <Section position="3" start_page="0" end_page="0" type="sub_section">
      <SectionTitle>
3.3 Evaluation on different inference modules
</SectionTitle>
      <Paragraph position="0"> In the following experiments, web pages were separately evaluated by focusing on one inference module based on the domain-specific ontology at a time. That is, the mixture weight is set to 1 for one inference module and the other is set to 0 in each evaluation. For comparison, the keyword-based VSM approach and the ontology-based system are also evaluated and shown in Figure 6. The precision and recall rates are used as the evaluation measures.</Paragraph>
      <Paragraph position="1"> And the ontology based approach means the combination of concept inference and axiom inference described in the section 3.2.</Paragraph>
    </Section>
  </Section>
class="xml-element"></Paper>