Android Malware Analysis

Published: 14 May 2026| Version 6 | DOI: 10.17632/bpkksc9v5s.6
Contributors:
,
,
,
,
,

Description

The datasets used in this study—CICAndMal2017, CICMalDroid2020, and Drebin—provide curated and structured information for Android malware analysis. All datasets are organized in CSV format and contain carefully selected static and dynamic features that support effective classification of benign applications and multiple malware samples. These features include behavioral indicators, permission-related attributes, and network or system-level characteristics, enabling comprehensive analysis for malware detection tasks. The datasets are preprocessed and formatted with clearly defined feature columns and corresponding class labels, making them suitable for machine learning and deep learning-based experimentation, benchmarking, and reproducible research in Android security. Additionally, the datasets are derived from publicly available Android malware repositories, where relevant features have been extracted and standardized into a unified format to ensure consistency and ease of analysis across all three datasets.

Files

Steps to reproduce

The CICAndMal2017, CICMalDroid2020, and Drebin datasets are provided in CSV format with standardized static and dynamic features for Android malware classification. To reproduce the study, load the datasets, apply preprocessing (cleaning, normalization, and feature selection), and use them for model training and evaluation under the same experimental setup.

Institutions

Categories

Machine Learning, Internet of Things, Explainable Artificial Intelligence, Android Malware, Artificial Intelligence of Things

Licence