Terminology Index

Glossary

2953 terms

Open concept maps
S
32

Statistical Analysis for Detection

Using statistics on SIEM data, counts, percentiles, baselines, anomalies, to write detections that catch unusual activity by behavior rather than known signatures.

Statistical analysis applied to SIEM data uses counts, ratios, percentiles, time-series math, and baselines to spot the unusual: a host's outbound bytes far above its normal, a sudden surge in failed logons, an unusual ratio of one event to another. As a SIEM-query-fundamentals topic, this approach extends signature-based detection with behavioral, baseline-aware logic written in query languages like KQL or SPL. It is a core building block of modern detection engineering.

Introduced in: SIEM Query Fundamentals

Examples

  • Detecting outbound traffic far above a host's baseline as anomalous.
  • Alerting on a surge in failed-logon counts via statistical thresholds.
  • Using percentiles in SIEM queries to flag unusual behavior.