MINING ROLES USING ATTRIBUTES OF PERMISSIONS

摘要

Recently, many approaches were proposed to generate roles using automatic techniques. However, most of these approaches generate many composite roles because they only optimize minimality of the state in role-based access control (RBAC). The responsibility of the composite roles is complex and hardly interpretable, which weakens the robustness of the RBAC state. In this paper, we propose to use operations and resources of permissions as the functional information in role mining algorithm and present a novel approach, role mining with functional features (FMiner), to reduce composite roles. The FMiner approach is a two-phase solution. Firstly, an initial RBAC state is built by formal concept analysis theory. Secondly, relative closeness is defined to measure the functional similarity between roles. We optimize the relative closeness and minimality of the initial RBAC state simultaneously. The experimental results demonstrate the effectiveness of the proposed approach on reducing composite roles.