Android Malware Detection: ML Pipeline

A machine learning pipeline that classifies Android apps using dynamic system-call features.

Repo

Tech Stack

PythonNumPyscikit-learnSVMLogistic RegressionRandom Forest

Highlights

  • Built a 360-feature pipeline and evaluated multiple classifiers on multiple datasets.
  • Designed a weighted-voting ensemble reaching 91.98% test accuracy with reproducible evaluation.
  • Focused on robustness and interpretability when combining models.