A Novel Algorithm for Differentiating Authorized Users from Fraudsters via Mobile Keypad Input Patterns During Password Updates

This study addresses the challenge of distinguishing legitimate users from imposters during password updates by developing the TKIP-RUB (Transforming Keypad Input Patterns to Recognize User Behavior) algorithm. A literature review revealed that existing algorithms, including the EPSB method, achieve limited accuracy in recognizing user behavior based on mobile keypad input. The research aimed to enhance authentication systems by answering the question: Can transforming historical input patterns improve the accuracy and reliability of user recognition? The hypothesis posited that the proposed algorithm would significantly outperform existing methods in accuracy and precision. To evaluate this, an experimental study was conducted using login attempts from 143 users who updated their passwords, resulting in a dataset of 629 records (486 training, 143 testing). The TKIP-RUB algorithm was integrated into a mobile authentication system to analyze user behavior and generate predictive patterns. Results showed that while the EPSB algorithm achieved an accuracy of 9.091%, the TKIP-RUB algorithm reached 53.147%, representing a fivefold improvement. This demonstrates the superior effectiveness of the TKIP-RUB algorithm in enhancing recognition rates, security, and positive predictive precision.This work underscores the importance of advanced algorithmic techniques in mitigating risks of unauthorized access and provides recommendations for future password creation strategies.