Anomaly Detection for Large Scale Analytics with Machine Learning
High velocity businesses need to analyze their data streams for the unexpected. Vast amounts of rapidly changing metrics and KPIs impact time-sensitive business decisions. Incomplete insights due to missed data anomalies can lead to lost revenue, dissatisfied customers, broken machinery or missed opportunities.
This white paper – part one of a three-part series – discusses design principles of creating an machine learning-based anomaly detection system. Topics include:
- Why companies need anomaly detection
- Types of machine learning
- What is an anomaly
- Design principles
- Timeliness
- Scale
- Rate of change
- Conciseness
- Definition of incidents
Offered courtesy of Anodot