Data on Vulnerability Detection in Android.

Area of Science:

• Computer Science
• Cybersecurity
• Machine Learning

Background:

• Android malware poses a significant threat to mobile users.
• Effective detection requires comprehensive datasets and advanced analytical techniques.

Purpose of the Study:

• To curate a diverse dataset of Android Package Kits (APKs) for malware research.
• To facilitate the development and evaluation of machine learning-based Android malware detection models.

Main Methods:

• Collected ~85,000 APKs from multiple sources including Google Playstore, Wandoujia, AMD, and Androzoo.
• Extracted static and dynamic features from APK files for analysis.
• Applied supervised machine learning algorithms for malware classification and feature selection.

Main Results:

• Developed a generic algorithm for effective feature subset selection.
• The dataset and methodologies support the creation of advanced Android malware detection systems.
• The findings contribute to ensembling classifiers for improved malware identification.

Conclusions:

• The curated dataset and analysis methods provide a valuable resource for Android malware research.
• The study highlights the efficacy of machine learning in detecting malicious Android applications.
• This work supports the advancement of cybersecurity measures for Android platforms.