| With the development of Internet, search engine has become the important measure to get information in our daily life.And information retrieval is the core of search engine. Due to the diversity and ambiguity of natural language, users'requirments is always not satisfied well in keywords search.And the breakthough of research on keywords search is also limited for this.With the emergence of semantic web, semantic-based information retrieval has become an effective way to improve retrieval ability. Ontologies, as one of major techniques in Semantic Web, its good conceptual level and logical reasoning support, and the ability to express semantic through relationships, can be adapted to assist the system to conprehand the users'query from the semantic level,and improve the accuracy. Vertical search is a search model which is advanced to direct towards the problems such as large amount of information, inaccurate queries raised by general search engine. Its emphasis is on "professional,exact, deep". Therefore, the accuracy requirements to query results is higher while the demand of professionalism in vertical search coincides with the characteristics of domain ontology. Found on the full research both at home and abroad, how to solve several problems in semantic search by using domain ontology and knowledge base is analyzed, and the main tasks and contributions of the thesis are as follows:(1)Domain ontology is the description to knowledge and features of specific areas.The composition of the domain ontology is highlighted and analyzed in this paper.OWL 2 ontology language and a variety of ontology construction method is introduced.Taking tourism as a backgroud, a travel ontology is built using the improved seven-step method.And OWL 2 is taken as the coding language.(2) Semantic query expansion is an important area of application of semantic technology. Eextending query terms by using domain ontology, and adjusting the query weight has a major impact on the query results.The query scope is analyzed and innovation is brought:the querywords can map to not only concept but also instances;the standard language and reasoning ablility of ontology is used to improve the query results.According to the different assosciation between query terms and extended terms, query weight is further adjusted. Meanwhile, taking tourism domain as the background, semantic query expansion is implemented and evaluated.(3)Semantic ranking is a key step in semantic search.The relevancy between instances and documents is taking important part in ranking.The algorithm which conbines syntax relevancy and semantic relevancy is put forward.Semantic relevancy is divided into equival relevancy and property relevancy. Information loss and ambiguity which is caused by diversity of natual language is avoided. At the same time, integrating semantic retrieval and traditional information retrieval avoids nothing return caused by the incompleteness of knowledge base system.As a ranking basis, the final document scores is obtained by syntax and semantics score integration. The effectiveness of ranking algorithm is also demonstrated.(4) While Semantic search in the present can not completely replace traditional search, some development in traditional information retrieval, can be applied to semantic retrieval. Drawing on the traditional information retrieval techniques, the tourism information retrieval system based on ontology is implemented, including documents dealing, knowledge management, semantic retrieval module.Mealwhile, the ranking tenology is applied, and the algorithm is evaluated that it can improve the effective of query and return more satisfied results. Through the system experiments, the performance evaluation, analysis, and optimization strategy is given. |
PDF Full Text Request