Achieving accurate anomaly detection requires more than statistics. Simple assumptions like normal distribution do not work in the real world. Time series data – representing anything from customer acquisition, to application performance, to manufacturing KPIs – tend to have many different behaviors that need to be modeled accurately. These include seasonal patterns, non-stationary behaviors, and intricate correlations between signals, among others.
Join Ira Cohen, Anodot’s Chief Data Scientist and Bill Vorhies, Editorial Director, Data Science Central, where they will discuss:
- Fundamental machine learning techniques for anomaly detection
- Requirements of an anomaly detection system in various use cases
- Issues and pitfalls to watch out for when implementing anomaly detection
- Common use cases and examples
Date: Thursday, October 13, 2016
Time: 09:00 AM Pacific Daylight Time
Duration: 1 hour
::UPDATE:: Click here to watch the full webinar.