Risk assessment based access control with text and behavior analysis for document management

Authors:

Abstract
In computerized systems, documents with sensitive information are generated, stored and accessed every day in large volumes. These documents are classified and disseminated only to appropriate personnel. Unintentional disclosure of sensitive information should be ultimately avoided. How to effectively provide access control of document disclosure is a key for secure business, government and military operations. Traditional access control is based on a simple rule, i.e., to test whether a user account that requests the access to information has been granted such an access. However, this design has been shown to provide no security guarantee due to emerging incidents including insider threats, account hacking, and human classification errors. In this paper, we propose a new access control mechanism based on a flexible decision design, which will not simply guarantee access to a document when a user account has been granted such an access, but comprehensively use text analysis and behavior analysis in a complementary way to quantify the risk of information disclosure and grant the access only if the risk is assessed low. Our evaluation based on notional documents demonstrates the effectiveness of this new access control design against erroneous document classification and malicious user behavior. The proposed access control mechanism shows potential to enhance the overall security in today's access control systems for document management.

This publication has 6 references indexed in Scilit: