|
|
--- |
|
|
language: |
|
|
- en |
|
|
license: cc-by-4.0 |
|
|
task_categories: |
|
|
- image-classification |
|
|
- tabular-classification |
|
|
tags: |
|
|
- ransomware |
|
|
- malware-detection |
|
|
- cybersecurity |
|
|
- pe-header |
|
|
- binary-analysis |
|
|
pretty_name: Ransomware PE Header Feature Dataset |
|
|
size_categories: |
|
|
- 1K<n<10K |
|
|
--- |
|
|
|
|
|
# Dataset Card for Ransomware PE Header Feature Dataset |
|
|
|
|
|
## Dataset Description |
|
|
|
|
|
### Dataset Summary |
|
|
|
|
|
This dataset contains PE header features (first 1024 bytes) from 2,157 Windows executable samples, comprising 1,134 legitimate software (goodware) and 1,023 ransomware samples across 25 ransomware families. Each sample is represented by numerical features extracted from the raw PE header. |
|
|
|
|
|
### Supported Tasks |
|
|
|
|
|
- **Binary Classification**: Distinguish between goodware and ransomware |
|
|
- **Multi-class Classification**: Identify specific ransomware families |
|
|
- **Malware Detection**: Train ML models for ransomware detection systems |
|
|
|
|
|
### Languages |
|
|
|
|
|
The dataset contains binary executable files; documentation is in English. |
|
|
|
|
|
## Dataset Structure |
|
|
|
|
|
### Data Instances |
|
|
|
|
|
Each row contains: |
|
|
- `ID`: Sample identifier |
|
|
- `filename`: Original filename or SHA-256 hash |
|
|
- `GR`: Target class (0=goodware, 1=ransomware) |
|
|
- `family`: Family ID (0 for goodware, 1-25 for ransomware families) |
|
|
- `0-1023`: Numerical features (byte values 0-255) from PE header |
|
|
|
|
|
### Data Fields |
|
|
|
|
|
- **ID**: Integer identifier (10000-11133 for goodware, 20000-21022 for ransomware) |
|
|
- **filename**: String (executable name for goodware, SHA-256 hash for ransomware) |
|
|
- **GR**: Binary integer (0 or 1) |
|
|
- **family**: Integer (0-25) |
|
|
- **0 through 1023**: Integer features (0-255 range) |
|
|
|
|
|
### Ransomware Families |
|
|
|
|
|
The dataset includes 25 ransomware families: Avaddon, Babuk, Blackmatter, Conti, Darkside, Dharma, Doppelpaymer, Exorcist, Gandcrab, Lockbit, Makop, Maze, Mountlocker, Nefilim, Netwalker, Phobos, Pysa, Ragnarok, RansomeXX, Revil, Ryuk, Stop, Thanos, Wastedlocker, and Zeppelin. |
|
|
|
|
|
## Dataset Creation |
|
|
|
|
|
### Source Data |
|
|
|
|
|
- **Ransomware**: Downloaded from VirusShare and Hybrid-Analysis databases |
|
|
- **Goodware**: Collected from PortableApps, Softonic, Windows 10 system files, and verified no-install software |
|
|
- **Verification**: All goodware scanned with ESET NOD32 Antivirus v15.1.12.0 |
|
|
|
|
|
### Annotation Process |
|
|
|
|
|
Ransomware samples were categorized using VirusTotal engine detections with three criteria: |
|
|
- At least 45 engines flagged as malicious |
|
|
- At least 15 engines flagged as ransomware |
|
|
- At least 10 engines nominated the same family |
|
|
|
|
|
## Considerations for Using the Data |
|
|
|
|
|
### Discussion of Biases |
|
|
|
|
|
- Dataset reflects ransomware families active during 2020-2021 period |
|
|
- Sample selection based on availability in public malware repositories |
|
|
- Goodware primarily consists of portable/no-install applications |
|
|
|
|
|
### Other Known Limitations |
|
|
|
|
|
- PE header features only (first 1024 bytes); no full binary analysis |
|
|
- File sizes range from 14.8 kB to 12.2 MB |
|
|
- Windows executables only |
|
|
|
|
|
## Additional Information |
|
|
|
|
|
### Dataset Curators |
|
|
|
|
|
Moreira, C.C., Moreira, D.C., Sales Jr., C. de S. de |
|
|
|
|
|
### Licensing Information |
|
|
|
|
|
CC BY 4.0 - You are free to share and adapt the material with appropriate attribution. |
|
|
|
|
|
### Citation Information |
|
|
|
|
|
|
