Measuring Vertex Centrality in Co-occurrence Graphs for Online Social Tag Recommendation

We present a social tag recommendation model for collaborative bookmarking systems. This model receives as input a bookmark of a web page or scientific publication, and automatically suggests a set of social tags useful for annotating the bookmarked document. Analysing and processing the bookmark textual contents document title, URL, abstract and descriptions we extract a set of keywords, forming a query that is launched against an index, and retrieves a number of similar tagged bookmarks. Afterwards, we take the social tags of these bookmarks, and build their global co-occurrence sub-graph. The tags (vertices) of this reduced graph that have the highest vertex centrality constitute our recommendations, which are finally ranked based on TF-IDF and personalisation based techniques.