Enhancing User Authentication: An Approach Utilizing Context-Based Fingerprinting With Random Forest Algorithm

Abstract

In the evolving world of Cyber Attacks, this research presents an innovative approach aimed at fortifying user authentication within the Application Programming Interfaces (APIs) Ecosystem. We employ a groundbreaking synthesis of Context-Based fingerprinting attributes alongside the Random Forest algorithm, assessing their efficacy in enhancing security in how the user is authenticated. The study emphasizes the model’s potential to advance the precision of identifying legitimate login attempts, positioning it as a superior alternative to conventional methods. A model evaluation investigates the capacity of the Random Forest algorithm, augmented with these attributes, to discern the authenticity of user login attempts. Results unveil an exceptionally accurate model with 99.5 % accuracy, showcasing elevated F1 scores, precision, recall, and a notable MCC. The paper’s insights underscore the Random Forest algorithm’s potential as a robust tool for user authentication using the user’s historical profile, significantly contributing to the domain of Cyber Security. As the digital landscape continues to evolve, this research endeavours to provide a pioneering solution, ensuring robust API security and endorsing the broader adoption of this groundbreaking model.

Keywords

• API ecosystem security
• API security
• context-based fingerprinting
• cyber security
• random forest
• user authentication