A Method Of Searching Entities Based On WordNet Noun Network

With the rapid development of the Semantic Web,there emerges more and more structured semantic data.Large amouts of structured data can be processed by machines to improve the experiences of users searching information on the Web.Data developers design and publish structured data to describe real world concepts and entities.This has arouse great needs of users to search entities of their interest conveniently in ease.With traditional web search technology,users can search for entities in Semantic Web.However,it hasn't taken advantage of the nature of structured data.Entities in Semantic Web are usually tagged by ontology classes contained in various ontologies.This brings obvious redundancy and heterogeneity arising puzzles for users to use the Semantic Web.In this thesis,a method to search entities with keywords is introduced.It provides category filtering tags to filter search results to improve the efficiency and experiences of searching.All of the category filtering tags are using WordNet noun network as the taxonomy.This thesis introduces methods to solve mainly two problems:a method to map ontology class hierarchies to WordNet noun network and implementing the entity search application Expedia,which enables users to use keywords to search entities and select interested category filtering tags to iteratively filter the search results.To compare and fill out.questionaires,an experiment is designed for test users to use different systems.With multiple statistic results,this thesis demonstrated that the system Expedia could obviously help users find entities with interest quicker and more conveniently.Both Expedia and other comparing systems are based on the same DBpedia 2014 entity dataset.