# Android Malware Category and Family Detection and Identification using Machine Learning

**Authors:** Ahmed Hashem El Fiky, Ayman El Shenawy, Mohamed Ashraf Madkour  
**arXiv:** [2107.01927](https://arxiv.org/abs/2107.01927)  
**Date:** 2021-07-05  
**Summary page:** https://www.summarizepaper.com/en/paper/2107.01927v1/

## Key points

- The paper addresses the growing threat of Android malware on the internet
- Highlighting the need for a better understanding of common Android malware categories and families
- Presenting two innovative machine-learning approaches for Dynamic Analysis of Android Malware
- One approach focuses on detecting and identifying Android Malware Categories
- The other approach targets Android Malware Families
- Developed by analyzing a substantial dataset containing 14 prominent malware categories and 180 prominent malware families from the CCCS-CIC-AndMal2020 dataset on Dynamic Layers
- Impressive accuracy rates: over 96% accuracy for Android Malware Category detection and over 99% accuracy for Android Malware Family detection
- Provides valuable insights into combating the escalating threat of Android malware and reduces analysis time
- Contributes to enhancing cybersecurity measures in today's digital landscape

## Summary

The paper titled "Android Malware Category and Family Detection and Identification using Machine Learning" by Ahmed Hashem El Fiky, Ayman El Shenawy, and Mohamed Ashraf Madkour addresses the growing threat of Android malware on the internet. The authors highlight the need for a better understanding of common Android malware categories and families in order to combat this pressing issue. They present two innovative machine-learning approaches for Dynamic Analysis of Android Malware - one focusing on detecting and identifying Android Malware Categories and the other targeting Android Malware Families. These approaches were developed by analyzing a substantial dataset containing 14 prominent malware categories and 180 prominent malware families from the CCCS-CIC-AndMal2020 dataset on Dynamic Layers. The results demonstrate impressive accuracy rates, with over 96% accuracy for Android Malware Category detection and over 99% accuracy for Android Malware Family detection. This not only provides valuable insights into combating the escalating threat of Android malware but also significantly reduces the time required to analyze smartphone malware. By leveraging machine learning techniques, this research contributes to enhancing cybersecurity measures in today's digital landscape.