Permission Based Android Malware Detection System

• Tech Stack: Python, scikit-learn, numpy, pandas, Android Studio, Java
• Github URL: Project Link

Developed a malware detection system based on minimum number of permissions required to classify the application as malware or benign with an overall 94.3% accuracy

Reduced number of permissions required for classification from 113 to 10 using model independent pruning techniques and dimensionality reduction

Classified malware applications into 45 malware families using K-means clustering algorithm and created an Android application based on the developed system for real time malware detection and classification