Research On And Implementation Of Metadata Similarity Calculation In The OAI-PMH

As a new kind of metadata interoperability protocol, OAI-PMH provides a useful solution for metadata transmission, sharing and dissemination. Because the protocol has the advantages of simplicity, low-threshold and platform crossing, it has been widely used in many fields. With the continuous expansion of the application for OAI-PMH, the metadata resources are showing fast-growing trend. In order to make users obtain the information they care about from rich metadata, it is urgent to find the relationship among metadata and get useful knowledge. As the foundation of finding information relationship and getting knowledge, similarity calculation has been widely applied in information retrieval, information recommendation and information mining, and it also plays an important role in information processing.Based on a deep study on similarity calculation methods at home and abroad, and combining with the structural characteristics of metadata, similarity calculation is introduced into OAI-PMH and a method which is used to calculate metadata similarity in the OAI-PMH based on VSM is proposed. Taking the impact of characteristic items in different position on the result of calculating weights in to account, we use the improved TF-IDF method by introducing position weight coefficient to calculate the weights of characteristic items after choosing out metadata characteristic items. We map metadata object to space vector, and then calculate metadata similarity by space vector angle. In this paper, we design and implement a Service Provider system and integrate metadata similarity calculation function into Service Provider in the OAI-PMH. Through the experimental results we can see that the method of calculating the similarity of metadata is effective and feasible, and we can get similar metadata from metadata sets. After calculating the similarity calculation among metadata in the OAI-PMH, similar metadata are provided to users when they are browsing and searching information that they are interested in. By this way, we can afford people much better appreciation service, help people getting the information they care about and greatly improve information service quality.