<?xml version="1.0" standalone="yes"?>
<Paper uid="E06-2016">
  <Title>The GOD model</Title>
  <Section position="2" start_page="0" end_page="147" type="intro">
    <SectionTitle>
1 Introduction
GOD (General Ontology Discovery) is an un-
</SectionTitle>
    <Paragraph position="0"> supervised system to extract semantic relations among domain specific entities and concepts from texts. Operationally, it acts as a search engine returning a set of true predicates regarding the query instead of the usual ranked list of relevant documents. Such predicates can be perceived as a set of semantic relations explaining the domain of the query, i.e. a set of binary predicated involving domain specific entities and concepts. Entities and concepts are referred to by domain specific terms, and the relations among them are expressed by the verbs of which they are arguments.</Paragraph>
    <Paragraph position="1"> To illustrate the functionality of the system, below we report an example for the query God.</Paragraph>
    <Paragraph position="2"> god: lord hear prayer god is creator god have mercy faith reverences god lord have mercy jesus_christ is god god banishing him god commanded israelites god was trinity abraham believed god god requires abraham god supply human_need god is holy noah obeyed god From a different perspective, GOD is first of all a general system for ontology learning from texts (Buitelaar et al., 2005). Likewise current state-of-the-art methodologies for non-hierarchical relation extraction it exploits shallow parsing techniques to identify syntactic patterns involving domain specific entities (Reinberger et al., 2004), and statistical association measures to detect relevant relations (Ciaramita et al., 2005). In contrast to them, it does not require any domain specific collection of texts, allowing the user to describe the domain of interest by simply typing short queries. This feature is of great advantage fromapracticalpointofview: itisobviouslymore easytoformulateshortqueriesthantocollecthuge amounts of domain specific texts.</Paragraph>
    <Paragraph position="3"> Even if, in principle, an ontology is supposed to represent a domain by a hierarchy of concepts and entities, in this paper we concentrate only on the non-hyrarchical relation extraction process. In addition, in this work we do not address the problem of associating synonyms to the same concept (e.g. god and lord in the example above).</Paragraph>
    <Paragraph position="4">  In this paper we just concentrate on describing our general framework for ontology learning, postponing the solution of the already mentioned problems. The good quality of the results and the well foundedness of the GOD framework motivate our future work.</Paragraph>
  </Section>
class="xml-element"></Paper>