Study On Social Tagging Recommendation Design Based On The Controlled Vocabularies In Medical Resources

Social tagging is a new kind of annotation method under Web 2.0 environment and tags are core elements in it. It affects the efficiency of information retrieval by tags’ diversity, vagueness, non-structured, semantic relations and other characteristics. The redundancy of the network information resources brings about inconvenience for users to retrieve fast and browse easily. Facing the bottleneck of freestyle annotation and folksonomy’s development, we turned to traditional taxonomy. Controlled vocabularies has a long history in the traditional taxonomy, and it classifies knowledge in part and is convenient to gather classified vocabularies with normalization and structural integrity according to the discipline of information resources. Therefore, to fuse the controlled vocabularies and social tagging together improves the efficiency and quality of information classification. Both to use controlled vocabularies extend user tag hierarchy, recommend for tags have effect on improving the retrieval efficiency and make use of mapping relationship between tags and controlled words. Thus it helps extend the thesaurus.This article analyzed the characteristics of tags of the resources and thesaurus, and built the mapping between them, which were based on the classification of medical thesaurus from Chinese Thesauri and the resources and tags getting from Douding as the system data source, in order to investigate the possibility of their fusion. According to the relationship between controlled words and tags, we built several matrixes such as co-occurrence matrixes, tag vector in the resource collection, co-occurrence matrixes between all tags from the collection and their similarity matrixes. It would get related information recommended to users that it was high frequency of clustering tags, which we took advantage of the statistical software to analyze. And we used the Visual Studio 2008, ASP.NET 2.0 program language and Microsoft SQL Server 2010 database management tools for system development and design.In this paper, a new kind of quad relationship as "user-category-tag-resource" was used. By the combination of tags and resources and users and the categories, we established the tagging mapping relationship with controlled vocabularies and extended the mapping through tagging clustering, in order to provide better tag recommendation mechanism and hierarchical browsing for users. Test proves that the proposed model based on the thesaurus of social tagging recommended model be better able to provide tag recommendation and classified navigation service for users.