Skip to Main Content
Studies show that a significant number of employees steal data when changing jobs. Insider attackers who have the authorization to access the best-kept secrets of organizations pose a great challenge for organizational security. Although increasing efforts have been spent on identifying insider attacks, little research concentrates on detecting data exfiltration activities. This paper proposes a model for identifying data exfiltration activities by insiders. It uses statistical methods to profile legitimate uses of file repositories by authorized users. By analyzing legitimate file repository access logs, user access profiles are created and can be employed to detect a large set of data exfiltration activities. The effectiveness of the proposed model was tested with file access histories from the subversion logs of the popular open source project KDE.