Cryptocurrency Malware Dataset

Published 2020. Recent years have seen dramatic growth in cryptocurrency trading, which has attracted malicious actors exploiting vulnerabilities to infect target systems.

Introduction

Threat actors deploy cryptocurrency malware to leverage infected devices for complex computational tasks. These threats are challenging to detect using manual or heuristic approaches, but machine learning based malware threat detection solutions have obtained promising results. The dataset targets MS Windows cryptocurrency malware threats, given the prevalence of this platform among cryptocurrency traders.

Dataset details

The collection contains over 700 samples across two categories:

• Malware samples – 500 real-world cryptocurrency malware files
• Benign samples – 200 legitimate files

All samples are executable files for the MS Windows OS platform, initially unpacked and decompiled using the Object-dump tool. Malware samples originated from VirusTotal and VirusShare, while benign samples came from the Microsoft Store and CoinMarketCap. The dataset includes 1,571 features per sample, creating dimensionality considerations for analysis.

Citation

Yazdinejad et al. (2020). Cryptocurrency malware hunting: A deep recurrent neural network approach. Applied Soft Computing, 96, 106630.

Download

The dataset is available on GitHub. For questions, contact ali@cybersciencelab.com.