SUP: Security User Profiles for Behavioral Data Platforms

• Main Authors: Sulayman, Iman I. M. Abu, Ouda, Abdelkader
• Format: Conference Proceeding
• Language: English
• Subjects: Analytical models; Authentication; Behavioral data analytics; Behavioral modeling; Computational modeling; Computers; Data analysis; Data models; Focusing; Information security; Security; Usability; User profile
• ISSN: 2768-0940
• DOI: 10.1109/ISNCC62547.2024.10759020

Integrating user behavior across domains to create comprehensive profiles is crucial for system efficiency and security. This paper delves into the use of data analytics techniques to gather user data from various application domains, enabling the creation of security user profiles that contain data suitable for user authentication. Specifically, these methods are capable of identifying patterns in the users' data and then using said patterns to detect legitimate but anomalous activities performed by the users. Furthermore, we propose an innovative mechanism that dynamically integrates these profiles to create a comprehensive and global security profile for user authentication. Typically, each profile represents a distinctive entity detailing infrequent user behaviors that significantly impact identity verification. The verification process involves posing user challenge questions generated from these integrated profiles. The generated user profiles ensure the selection of questions that meet both security and usability requirements. We uphold security by issuing each question from the integrated user profiles only once, protecting user responses from potential compromise. We prioritize usability by using fresh data, which helps legitimate users recall and complete the challenge easily.