Entity And Relationship Acquisition Of A Domain Ontology

With the popularization of the Internet, information on the Internet growing explosively. In order to cope with the information explosion, we need some kind of automated tools to help people find the information they really need from a mass of electronic documents. It is in this context, research of information extraction born out.Information extraction refers to the automatic extraction of the specific factual information such as entities, relationships between entities, and attributes describing entities from unstructured sources. An Ontology is an explicit specification of a conceptualization, when the ontology is used as the background knowledge of IE system, it can guide information extraction process and improve extraction performance.In this paper, the domain is limited to the field of cloud computing industry, and information extraction task includes the identifying of entities, such as company names and related attributes, as well as relationship extraction, such as relationship between one company entities and its property, relationship between companies. The work of this paper is summarized as the following three aspects:1. After learning the basic knowledge of ontology and ontology establishment method, based on the news related to cloud computing industry, hand-established an ontology of the cloud computing industry and a rule base; and implemented a mapping from natural language query to SPARQL query.2. Rules-based methods and statistical methods were used for the identification of entities of different types. In particular, to identify the entity of company name, proposed a conditional random field based and twice-scan method, and achieved good experimental results.3. Proposed an ontology rules based method in identifying the relationship between entities, which using the rules of the relationship between entities of the ontology, to correct recognition results gained from previous relationship extraction.