| With the development of information technology, a great deal of digital information is accumulated. OAI-PMH, as a simple and flexible protocol to solve interoperability and sharing of mass digital information, has drawn more and more people's attention. However, most of service providers, which have been implemented, only make a simple index for the harvested metadata without further organization. When the harvested metadata becomes massive, it will bring many problems such as the difficulty in data organization reasonably, long cycle and slow speed in data searching. Thus, how to effectively organize the metadata harvested from service provider to facilitate user to search, browse and manage metadata and to achieve efficient metadata interoperability and resource sharing becomes a major topic which desiderates to be solved currently.On the basis of in-depth study of the OAI-PMH protocol, this paper analyses and compares the deficiency of existing index and classification methods, and proposes a clustering method to organize the mass metadata which is harvested by service provider of OAI-PMH. This provides a new effective way for the rational organization of the metadata in the service provider. Because FIHC algorithm is based on frequent item to clustering, and doesn't consider the words of semantic information, it is difficult to improve its accuracy further. By clustering method, in particular the discussion of FIHC algorithm, this paper proposes a FIHCSS algorithm for the defect of FIHC. FIHCSS algorithm has been improved by the word similarity computing based on How-net. The score function has redesigned by adding the words semantic information. The experimental study shows that the FIHCSS algorithm not only inherits high efficiency, rapid speed, good scalability, and a cluster tree as the results of clustering from FIHC algorithm, but also greatly improves clustering accuracy.Therefore, this paper proposes the OAI-PMH service provider framework based on FIHCSS algorithm, which applies the FIHCSS clustering algorithm to the OAI-PMH for organizing the metadata harvested by service provider. According to the design of database, use case, class and system architecture, OAI-PMH service provider system based on the FIHCSS is implemented. Experiments show that this method can well organizes a large quantity of metadata that service provider harvests back, greatly improves the search speed and facilitates users to browse metadata by category, it is more suitable to organize metadata that service provider harvests. |
