摘要：Role mining for Role-Based Access Control (RBAC) has emerged as a challenging area of information security. Current role mining models are mainly based on data mining with RBAC semantics and are not the pure mathematical models. Almost every model need a specific algorithm and the expressions of results and process are not vivid and explicit enough. To address these problems, this study proposes a role mining model called network-clique-finding (NCF) model, which is based on cliques finding of network. By this model, the role mining problem can be mapped onto the classic problems of graph theory and can be solved by using many well-known algorithms and tools. The correctness of NCF model is verified both theoretically and experimentally. Furthermore, some experimental issues is discussed and based on the experiment, our study is compared with some other methods. Finally, we conclude our work
