Relational Database Keyword Search Based On E-R Diagram

With the development of information retrieval technology and relational database technology, how to apply information retrieval techniques to the relational database, to obtain useful information, when general users or Web users do not need to know the relational database schema, and do not need understand Structured Query language syntax. Only using the search engine, relying solely on a simple search keywords from a relational database to obtain the information needed to become a recent research focus of many researchers. In this paper, on the basis of many research achievements of keyword search, we focus on the search effectiveness, search efficiency and results returned, and we conducted a more in-depth study of the problems in the existing search system, and the solutions approach.First of all, we introduce the knowledge of keyword search on relational database, and the two main directions of the research which is based on the database schema graph and database data graph were described, and we analyze their advantages and disadvantages.Secondly, in order to solve the problem of the keywords search on relational database systems, which still exist in the search effectiveness is not accurate, and the search efficiency is not high,we propose a new type of search system framework based on the E-R diagram of the relational database keyword search. According to the significance of the different entities in the E-R diagram, entities are grouped. Using keyword index to determine the collection of entities of the search When we search keyword, to narrow the search range and improve query efficiency. And through the new mode schema graph based search algorithm to improve search effectiveness and search efficiency.Again, for results returned of the current keyword search on relational database systems technology which can not be satisfied for user, according to form generation mechanism, first search system converts results to a class of natural language form which is returned to the user, the user will be easier to select their own inquiries with the same purpose of the form, then search system presents the specific result to the user, and the user can select the detailed data which they need. Finally, the new keyword search methods are verified with the experiment on the real data, and the methods which is effectiveness,feasibility and superiority are proved.