Download
AnomalyDetection is an R package developed by Twitter for detecting anomalies in seasonal univariate time series. It implements the Seasonal Hybrid Extreme Studentized Deviate (S‑H‑ESD) test, which reliably identifies both global and local outliers in data with trends and seasonality—commonly applied to system metrics, engagement data, and business KPIs.
Features
- Detects anomalies robustly even under seasonality and trend
- Implements S‑H‑ESD: hybrid ESd with seasonal decomposition
- R package easy to integrate into analytics workflows
- Scalable for large series or batch processing
- Produces clear diagnostic output and flags
- Used extensively for system and product telemetry analysis
Project Samples
Categories
SystemLicense
GNU General Public License version 3.0 (GPLv3)Follow AnomalyDetection
Other Useful Business Software
Award-Winning Medical Office Software Designed for Your Specialty
RXNT is an ambulatory healthcare technology pioneer that empowers medical practices and healthcare organizations to succeed and scale through innovative, data-backed, AI-powered software.
Rate This Project
Login To Rate This Project
User Reviews
Be the first to post a review of AnomalyDetection!
