An Automated Domain-Specific Answer Ontology Construction

Abstract

Recently, with the fast development of interactive information sharing on the internet, the query-oriented answer search services are more and more popular, such as Yahoo Answers, Google Answers and so on. Their common advantages are high precision, domain-specific search, clear format, etc. However towards domain-specific search, people usually are unable to determine suitable concept terms as queries to submit on account of their lack of domain knowledge. In this paper, we propose an approach for constructing a domain-specific answer ontology automatically in respect of Chinese queries to solve the said problem. First, queries and their answers are collected from a web search space. Second, for extracting implicated concept terms from collected queries and answers, the CKIP system is utilized to make a Chinese part-of-speech tagging procedure to segment. Thirdly, use a similarity measure to converge duplicate queries with their corresponding answers and take fuzzy clustering method and degree of membership to describe the relationships between converged queries and their corresponding answers. Finally, to generate a domain-specific ontology is based on an improved hierarchical agglomerative clustering algorithm to hierarchically group queries.