Personalized Semantic Retrieval System In The Field Of Science And Technology Policy Research

With the continuous development of national science and technology level, science and technology policies has gradually become an important part of the national policy system.The research of science and technology policy has drawn our academic and political circles, and many other scholars’ concern and attention.For science and technology policy researchers, in a number of policy documents quickly locate the information they are interested in is that they are very concerned about. Aimed at the present situation of information retrieval technology is ignored on semantic information and user information considerations have led to the recall ratio and precision rate is low. In this paper, relying on the "Hebei province Science and Technology Policy Data Management Service Platform" project system, based on the research of science and technology policy domain ontology semantic semantic association and user personalized information retrieval, the main contents include:(1)The construction of science and technology policy domain ontology. This article is based on the research in the field of science and technology policy, based on thesaurus documents that issued by the Chinese Academy of Sciences, analysis of the characteristics of science and technology policy and to build science and technology policy domain ontology with reference to the seven-step method.(2)In-depth study of information retrieval and knowledge and technology related to ontology, the concept of semantic extension model based on domain ontology is put forward. In order to improve the result recall of information retrieval, using ontology concepts and the relationships between the concepts to expand user semantic query.(3)In-depth study Lucene search engine and the knowledge and technology personalized related to information retrieval, based on user information and query logs, using LDA theme model establish user interest model, get user concern object, and optimizing Lucene documents calculation methods based on user interest-word, designing personalized information retrieval system based on user interest. And then gives the process of user interest model building and the retrieve sorting algorithm optimization.(4)To the "Hebei province Science and Technology Policy Data Management Service Platform" project system for relies on, increase its full-text search function and for optimization. Using JSP language to build personalized information retrieval module based on Lucene open source framework. Through reptile to collected science and technology policy information, organize stored in the database, and then build an index for query. Calculating the document similarity based on the search terms of semantic extension and user interest model, making the system understand the semantic information of user’s query better, increasing the documents score which user interest in, and thereby affecting the ranking of that documents. Finally, using experiments to verify the validity of this method. It can improve the recall rate and precision of search function as well as achieve the personalized retrieval enhanced satisfaction of user queries.