Instant User Insight: Dynamic Activity Profiling
Mehdi Mekni, Charles R Barone · 2024
Continuous authentication ensures ongoing verification of a user’s identity during an active session, removing the necessity for one-time logins. Real-time user activity fingerprinting, an effective method within continuous authentication, examines behaviors like keystroke patterns, mouse movements, and browsing habits to craft a distinctive user profile or "fingerprint." This approach boasts advantages over traditional authentication methods such as passwords or tokens, enhancing security by complicating imitation of user behavior while also streamlining the login process. This paper introduces Gargoyle, an intelligent software solution that we design, develop, and integrate. Additionally, we present a comprehensive software development methodology covering requirement analysis, software architecture, and design. Gargoyle employs a machine learning (ML) model that we develop, train, and validate. This ML model dynamically evaluates the smart continuous authentication system’s performance, assessing error rates and accuracy factors.