Approximate Query Method Based On Relational Database Keyword Semantic Research

Along with the universal application of Internet and the rapid expansion of web data, the database of online keyword search approach has become a major means of people online consulting and searching the information. On the diversity of the data organization form and structure, however, researchers have proposed different data models(hierarchical, network, relational, object-oriented, semi-structured data model), and then it also leads to the development of many new data management technology is in constant(such as XML data management, data flow management, Web data integration, data analysis and mining, etc.). Relational database because of its data structure is simple, clear, and has higher independence and better security and confidentiality, has become the most widely used universal database and as an important part of Deep Web. Therefore, relational database retrieval technology should have a higher user friendliness and versatility. Database management system to provide complex database tools, can make the professional users using SQL realized retrieval, but for most ordinary users, like Google, Yahoo and other search engines by submitting a few keywords that can retrieve relevant information keyword query method, is more likely to be its acceptance and use. However, due to the differences in the expression of non-professional users query intentions and retrieval of concern and understanding of relation database stores the content limited, caused many database information cannot be obtained, the decline in the level of user satisfaction and tentative retrieval times will increase. So against retrieval methods, means and technology need to constantly improve, make the associated with the keyword semantic(but not explicitly given keyword) contents of the query result is given, improved the precision, and better provide simple and reliable access to information services.At first, this paper puts forward an assessment of the semantic relevance between all the data stored in the database and the user for a given keyword query condition, that is, the method of analyzing the Coupling relationship between terms-TCR(Term Coupling Relationships). According to the TCR Algorithm evaluated coupling relationship between all the data of keyword database and users in the initial conditions(the two part including the coupling and coupling), and to create the order(i.e., ranking list) of all terms in database for each query keyword according to their coupling relationships to the query keyword. Then using ranking mechanism gets in database Top-k terms as candidate keyword are recommended to the user, the user through reconstruction of their intentions of most close to the conditions offered to re-query retrieval system. The query reconstruction system chooses DISCOVER system based on candidate network to re-query processing, that is, the TCR method is integrated into the traditional query system. Embedded TCR, which is the method of retrieval system can make the lack of experience of the user based on the system analysis of the initial conditions and recommend a high correlation between keywords, allowing users to reconstruct query condition to conform to his needs, in a sense, realized the purpose of the initial approximate query. And which allows users to re-build their subjective and select keywords, make the retrieval more humane, more can satisfy the user’s real thoughts. Combined with the advantages of accurate matching, the retrieval efficiency and user satisfaction reached higher requirements.The TCR algorithm is applied to analyze the the keywords and the database of data coupling relationship using the DBLP, movielens and IMDB data set in the experiments, which proves that the parameter α caused the different databases for the different effects. And the recommended keywords were returned by the Top-k recommended algorithm. Further, in the DISCOVER system to achieve the final query keywords.