Firewall Metrics for Unsupervised Intrusion Detection Research

Aggregating firewall metrics used to train and test unsupervised machine learning-based Intrusion Detection Systems for detecting malicious active scans on corporate networks. It was created by Matana da Rocha, Paulo and used in the associated research paper, with a last update recorded in February 2026.

Use Cases

• Train unsupervised models on firewall metrics to detect patterns indicative of malicious active scanning.
• Benchmark anomaly detection algorithms using the specific firewall metrics collected for IDS research.
• Analyze temporal patterns in firewall logs to identify characteristics of network scanning behavior.

Strengths

• Dataset is directly tied to a published research paper on ML-based IDS.
• Last update was recorded in February 2026, indicating recent curation.

Limitations

• Specific data volume, features, and sample data are unavailable for assessment.
• The dataset's scope is explicitly focused on active scan detection, limiting generalizability to other attack types.

Provenance

Source

Harvard Dataverse

Collection Method

Firewall logs from corporate networks, as described in the associated research paper.

Freshness

Last updated 2026-02-05.