Research On Keyword Query In Personal Dataspace Management System

With the development of computer technology, personal data gradually presents characteristics of massive, complex and heterogeneous. Traditional data management technology cannot meet the needs in this new situation. Dataspace, as a new data management technology, can meet the needs of the vast amounts of heterogeneous data management and service management needs because of its pay-as-you-go integration method and loose data model.Keyword search becomes a research hotspot in data retrieval field in recent years. Compared with query language, SQL and XQuery, keyword search has its unique advantages, which is very user-friendly to user without learning complex syntax specification and understanding the mode structure of system. This paper studies the keyword search mechanism in personal data space management system. The main contributions can be summarized as follows:(l)Analyzed the research actuality of personal dataspace management and existing keyword technology, we propose a new keyword querying method based on semantic. In this method, the node in data graph is analogous to the entity, attribute or attribute value in the real-word, grouping node matching keyword makes keywords with semantic information. Linking the keyword with specific environment and clearing the keyword semantics can improve the accuracy of matching user's query intention.(2)A new method of keyword classification is proposed. Solve defects of traditional method of classification. This method classifies keyword as predicate keyword which indicates the scope of query and result keyword which helps to show the user's query result. Keyword classification enriches the semantic information and this method can display query result more friendly. In addition, when only exist the predicate keywords cannot clear the user's query intention, we present a reasonable closeness function to rank the target node, in order to improve the effectiveness of query results.The experimental results demonstrate that the semantic-based keyword search method is more effective than traditional query methods to express the user's query needs and have great improvement in the effectiveness and efficiency of query.