Social linkage and ranking model for tags-based resources

Abstract

With the proliferation of social media, it is becoming important to support a significant amount of user tags in selecting the most appropriate resource description during the search process. In this paper, we propose to identify and structure the links between resources by taking into account a resource social dimension. Each resource is assigned to a cluster of tags hierarchy. The clusters of tags are formed by a classification method while the hierarchical classification of tags within clusters is defined using a hierarchy classification algorithm. User’s query is expanded by a social dimension and the clusters of tags are used to facilitate the search and ranking process. The results of our experiment, crawled from Delicious Folksonomy, demonstrate significant improvement over traditional retrieval approaches.