Security Defect Identification of Android Applications by Permission Extraction using Machine Learning

Security Defect Identification of Android Applications by Permission Extraction using Machine Learning

연구 분야: Analysis

학회: International Conference on Artificial Intelligence of Things

The Android platform security is at danger due to malicious applications. Due to the quantity and variety of these applications, traditional solutions are losing their effectiveness, making Android smart phones frequently vulnerable. Features are extracted during static analysis rather than running any code. Static analysis is more effective in general. Features taken from the manifest file can be used in static analysis of an Android application. The proposed model uses static analysis for permission extraction of permission from the application having security defects. A dataset of 1058 applications from the real world, 578 of which are benign and 480 of which are malicious, has been created. The model has been evaluated to ensure the effectiveness of the static analysis method for confirming the security of Android applications using machine learning techniques.