Research on authority users tag recommendation based on divided topics

Currently, researches on the quality of social tagging, including problems triggered by synonymous tags, lack of hierarchy and standard structure, have become a hot issue in the academia. Based on the construction of resource-tag network, this paper uses CPM to identify users´ interests and classify them into different topics. Then, user-resource network is constructed, and the key degree of each node is calculated. On basis of these, authority users in each topic are selected, and the most frequently used tags by these authority users are recommended to common users, in order to improve the quality of user tagging.