A Deep Dive into Relevant Feature Identification for Unveiling PII Leakage in Smartphones

Advancement in networking technologies has led to an increase in mobile phone usage exponentially, simplifying people's lives. However, this surge has created pathways for unauthorized access to personal information. These devices store large amounts of sensitive data, raising concerns about data leakage or exfiltration. Therefore, detecting unauthorized data leakage from these devices becomes critical. To address this, we propose enhanced machine learning frameworks aimed at accurately detecting personal information exposure. Our framework integrates the XAI algorithm, LIME, to enhance inter-pretability and comprehension of the detection process. Through experiments, we demonstrate that incorporating LIME leads to improved explanations and responsiveness, offering a more user-friendly and transparent solution.