Evaluating Password Security and User Awareness Using Machine Learning to Prevent Cyber Risks

Abstract

Cyber-attacks on the rise, protecting sensitive information through robust password security is more important now than ever. While multi-factor authentication and other advanced security technologies have become more prevalent, the password remains the most widespread form of access control. But weak, and guessed passwords remain a significant vulnerability. This article's objective is to investigate the extent machine learning could help improve password security and educate users, we focus on building a strong model “use when we apply passDefenderX to detect weak passwords and mitigate cyber risks" The main aim of this research is to test the effectiveness and performance of several types of machine learning techniques for password strength estimation, and the extent to which user knowledge on weak passwords can be refined with data-driven methods. It investigates several ML techniques including XGBoost, SVM, Naive Bayes and a newly proposed PassDefenderX model for predicting the success of passes in tennis on different accuracy performance metrics (e.g. accuracy; precision; recall). Performance evaluation results or PassDefenderX is compared against other approaches and concludes that our approach surpasses all others and show the maximum values in terms of the accuracy, precision with 0.9715, precision with 0.9713, recall 0.9715 respectively which makes it a better solution to detect insecure passwords brewing any cyber threat bursts. Results showed that 21 PassDefenderX Model significantly increase the security of passwords by identifying weak Fig. With its – overall good – consistent performance in all metrics, the system proves a reliable tool for identifying password weaknesses. Results of this study underline the potential of machine learning as a solution to ongoing issues with password security and user behavior, providing constructive models for enhancing cybersecurity.