Unveiling Moblie Money Phishing Scam Employing Reinforcement Learning Strategies

computer science

This abstract explores the widespread use of mobile money in regions lacking traditional banking infrastructure. Despite its efficiency, the technology faces challenges from malicious actors who exploit social engineering for scams and frauds in the absence of robust security measures. Addressing this gap, the paper introduces a fresh approach employing reinforcement learning techniques like Q-learning and Markov decision processes, along with deep reinforcement learning algorithms such as deep Q-learningThe proposed method aims to create models that understands and counteract phishing attacks by identifying optimal sequences of attacker actions through reinforcement learning and deep reinforcement methods. Real-world attack scenarios encountered at Orange and MTN telecoms are used for experimentation, comparing the effectiveness of reinforcement learning and deep reinforcement learning algorithms. RL exhibited better learning performance compared to DRL.Q-learning was found to have superior learning quality and faster execution time than certain DRL algorithms. Also, Some DRL algorithms were identified as beneficial in improving the understanding of scammer-victim interactions during mobile payments..

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.