Study Of Cross-Database Search Methods Based On Metadata

With the explosive growth of network resources, particularly in the library and information field, type and number of database resources is increasing. Because each database resource has its interfaces, site structure, authentication, users often need to get familiar to the database from scratch. Different database contents are often complementary, so users often search the data from one database to another. To solve these problems, cross-database search systems come into being.Cross-database search systems often use B/S model, providing users with a unified search interface, users enter search request in this search portal, the system will retrieve data from each database simultaneously, after being grouped and sorted, the search results will be presented to the user. Now mature cross-database retrieval systems are mainly based on meta search engine, data integration and middleware technologies. Some of these systems have low efficiency; some do not open data sharing interface; some systems'configuration is complex and difficult to maintain; some data description is non-standard and difficult to exchange data. To solve the above drawbacks, Cross-database search system model based on the meta-data warehouse is proposed.Metadata is data using to describe data, uses to describe specific resource object, and can locate and manage the object, and help it find and access data. DC Core metadata attracts extensive research interests in various areas from the beginning. After years of research and development, DC Core metadata has been set as basic metadata of description of resource objects in this field by a number of organizations of many countries, China Digital Library Standards Development project will also target DC as the basic element of digital resources data sets and will be expanded for different types of characteristics of the literature.Cross-database searching systems include many types of resources, such as journals, conference papers, dissertations, electronic books, newspapers, etc. Therefore, we must fully take the similarities and differences of types of resources into account, and also we must taking into account that the main function of cross-database retrieval systems are not store, share data but retrieve data, unified search large and complex data of distributed and heterogeneous system. So we divide the using metadata into three levels, including the core elements, the core elements of resource type, retrieved auxiliary elements.Metadata apply the important content of profile model theory, that is, when designing resource metadata structure, we must try to search the corresponding entities of elements needed to be described and managed from general metadata specification, construct our own meta-data specification framework by using these entities. When constructing cross-database searching systems metadata framework, this project fully learns the Chinese Digital Library Standards Development Project on the definition of metadata specification of all kind of literature and detailed analysis the characteristics of the commonly used journal articles, dissertations, e-books these three documents and gives a detailed definition of the metadata specification.Finally, cross-database retrieval system framework based on the metadata storage will be discussed, and focused on database design and search engine theses two core issues, finally the system is tested.