Malware Analysis Datasets: PE Section Headers for Static Detection

Static analysis data from PE section headers, extracted from Cuckoo Sandbox reports. The dataset contains examples of malware from VirusShare and goodware from portableapps.com and Windows 7 directories. It was created as part of PhD research on malware detection and classification using Deep Learning.

Use Cases

• Training binary classifiers to distinguish malware from goodware based on section header features.
• Analyzing the relationship between section entropy and maliciousness.
• Studying differences in memory allocation (virtualsize) versus disk size (sizeofdata) between malware and benign software.
• Developing feature sets for deep learning models focused on PE file structure.

Strengths

• Includes a clear binary classification label (malware/goodware).
• Features are derived from a known analysis tool (Cuckoo Sandbox).
• Sources for malware (VirusShare) and goodware (portableapps.com, Windows 7) are specified.

Limitations

• The total number of rows and examples is unknown.
• The description lacks column-level documentation beyond the listed features.
• The last update date is unknown, which may affect relevance against newer malware.

Provenance

Source

VirusShare (malware), portableapps.com and Windows 7 directories (goodware).

Collection Method

Static analysis data extracted from 'pe_sections' elements of Cuckoo Sandbox reports.