<?xml version="1.0" standalone="yes"?>
<Paper uid="C02-2001">
  <Title>A DAML+OIL-Compliant Chinese Lexical Ontology</Title>
  <Section position="1" start_page="0" end_page="0" type="abstr">
    <SectionTitle>
Abstract
</SectionTitle>
    <Paragraph position="0"> This paper presents an ongoing task that will construct a DAML+OIL-compliant Chinese Lexical Ontology. The ontology mainly comprises three components: a hierarchical taxonomy consisting of a set of concepts and a set of relations describing the relationships among the concepts, a set of lexical entries associated with the concepts and relations, and a set of axioms describing the constraints on the ontology. It currently contains 1,075 concepts, 65,961 lexical entries associated with the concepts, 299 relations among the concepts excluding the hypernym and hyponym relations, 27,004 relations between the lexical entries and the concepts, and 79,723 relations associating the lexical entries with the concepts.</Paragraph>
    <Paragraph position="1"> Introduction The Semantic Web relies heavily on formal ontologies to structure data for comprehensive and transportable machine understanding [Maedche and Staab 2001]. Therefore, constructing applicable ontologies influences the success of Semantic Web largely. An ontology mainly consists of a set of concepts and a set of relations that describe relationships among the concepts. An upper ontology is limited to concepts that are abstract, generic, domain-broad, and articulate. Cycorp constructed an upper ontology - Upper Cyc(r) Ontology. It consists of approximately 3,000 terms, i.e. concepts and relations. It has been used for organizing the upper structure of a knowledge base - the Cyc(r) KB. A working group of IEEE (P1600.1) is also trying to standardize the specification of the upper ontology. An upper ontology, called IEEE SUO (Standard Upper Ontology), is expected to enable computers to utilize it for applications, such as natural language understanding and generation, information retrieval and extraction, Semantic Web services [McIlraith et al. 2001], etc. It is estimated to contain between 1,000 and 2,500 terms plus roughly ten definitional statements for each term.</Paragraph>
    <Paragraph position="2"> This paper presents an ongoing task that will construct an upper-ontology-like ontology for Chinese research and applications. We refer to it as CLO (Chinese Lexical Ontology). In addition to the structural portion, the CLO will contain Chinese lexicons associated with the concepts and relations. A pure ontology containing concepts only (without lexicons) is abstract. A lexicon-associated ontology makes the substantiation of abstract concepts easier.</Paragraph>
    <Paragraph position="3"> HowNet defines 65,961 Simplified Chinese lexical entries by a set of predefined features including 6 categories of primary features and 100 secondary features, and several symbols, in which the primary features are in a taxonomy with single inheritance. The taxonomy is essentially regarded as the taxonomy of the CLO. However, the Chinese lexical entries defined in HowNet are simplified version. They are not suitable for Traditional Chinese research and applications. A traditional version of Chinese dictionary released by Sinica of R.O.C. is frequently used for Traditional Chinese NLP. By combining the Traditional Chinese dictionary and the HowNet, we attempt to construct the CLO and represent it in the semantic markup language DAML+OIL since DAML+OIL is currently a basis of Web ontology language.</Paragraph>
    <Paragraph position="4"> The task of constructing the CLO can be divided into three portions. Firstly, a hierarchical taxonomy of concepts including relations among the concepts is required. In our case, we utilize the hierarchical primary features of HowNet to form the structure. Secondly, a set of lexical entries should be associated with the concepts and relations. Thirdly, a set of axioms that describe additional constraints on the ontology are required. This paper addresses the ongoing construction task and a brief introduction of Web ontology languages.</Paragraph>
  </Section>
class="xml-element"></Paper>