A framework of service oriented semantic search engine

A service oriented semantic search engine framework is proposed to extract automatically and accurately information on service provider on the Internet. Domain knowledge on theme and abstract service provider are modeled by ontology to guide information extraction. Theme identification for each web page is executed after the page is fetched by search engine. A page or a part of content of a page is assigned a theme. Concrete service provider entity and attributes are extracted from content of web pages based on themes of the pages. Several algorithms are developed. Computational experiments demonstrate that service oriented semantic search engine shows high recall rate and precision. Highest recall rate of concrete service provider entity exceeds 97% on tested web sites. Precision of concrete service provider entity reaches up to 100% on two tested web sites. It also shows good recall rate and precision of concrete service provider attributes.