Domain ontology based document summarization and query processing methodology

Abstract

Even though information-oriented society is coming and lots of documents are being created rapidly, we couldn’t meet a demand due to lacks of human resource and time. Over last half a century, previous research has tried to solve the problem by not only NLP but also other methodologies. But it looks like difficult to solve that due to language’s complexity and anomalism. Especially, they have been in troubles in understanding contents as well as restructuring key contents precisely. To solve these problems, we use an ontology. The ontology can make a system easy to restructure sentences and raise readability up by giving the system common format. An expert defines domain-ontology which is specification what contents should be summarized. After initial work, the system extracts key concepts and a relationship among concepts based on domain ontology and summarizes useful information to a user. A semantic model is created from an original document by syntactic and semantic analysis and it is transformed into sentences by rule based translation process. Besides, the system analyzes a query based on domain ontology. It makes the user possible to get an extended result.