Research On Data-Community Retrieval Method Over Relational Databases

Nowadays, with the rapid development of the Internet, the keyword search technology provides key technique support to the web search and things like that, which is the key of web search to achieve great success, so that users can use their own familiar simple keywords to search for the information they need on the Internet. In many databases, relational database is one of the most popular and most used databases, and as long as the users master some knowledge of basic mode of structured query language and database, they will be able to carry out the operation. However, the vast majority of netizens do not have these specialized knowledge, so how to make the relational database support the efficient keyword search has become the focus of research recently, because relational database keywords can help users get away from the structured query language (For example SQL).Compared with network search engines, relational database keyword search has different characteristics, such as:database is composed of tuples; there are key value or semantic relation between tuples and so on, so a corresponding retrieval should be carried out to achieve retrieval when the keyword search is applied to relational database, to make the retrieval results have further information integration to some degree and get the object that the user really need.The data community retrieval method of the relational database mentioned in this paper is based on the center node as the drive and the combination of each keyword nodes that it can reach to determine a community. This paper introduces a community composed of nodes and edges, and nodes are divided into three categories:keyword nodes, the center node and the path node; edges can either be undirected edges or directed edges. With the public nodes of each keyword can reach by extending a certain distance as the center nodes, there are information of each keyword nodes that the center nodes can reach in the center node structure, according to which to determine whether it can meet the condition of the formation of a community, namely the information that users want to get. As for the sequencing problems, some corresponding standard adjustment can be carried out by using the overhead from the central node to each keyword nodes as the basis, in order to get the optimal result. This paper adopted the data community retrieval method framework of the relational database designed according to the method mentioned above and brought about related algorithm. With the DBLP data sets, the corresponding algorithm are verified, whose structure confirmed the feasibility of algorithm thought and the efficiency of specific algorithm. The retrieval method thought in this paper can not only be applied to the retrieval of the data community, it can also be extended to other areas, compared with keyword retrieval results of traditional relation database, this algorithm effectively improves the defects of single results containing less information and lacking corresponding dependence relationship between the results, and through this paper it makes the single result contain relatively abundant information and the users can have a clearer and more perspicuous understanding of the relationship between retrieval results.