Time Series Anomaly Detection: Optimizing your Machine Learning Jobs in Elasticsearch

Rich Collier

Solutions Architect, Machine Learning Specialist

Elastic

Tom Grabowski

Principal Product Manager

Elastic

Got logs and metrics (or any time stamped data really) in Elasticsearch, and looking to craft your 1st machine learning job to automate anomaly detection? Watch a practical walkthrough on efficiently progressing your machine learning projects from concept to production.

Elastic Machine Learning experts kick off the presentation with the Elasticsearch and machine learning basics, and then demomstrate the iterative process of crafting and optimizing your machine learning jobs.

Highlights include:

• Leveraging Kibana visualization tools to build intuition about the data
• Using scheduled events to exclude expected anomalies from data / model
• Forecasting into the future using the trained model
• Getting proactive via alerting integrations

Additional Resources (Related content)

• Blog posts:
  
  • Scheduled Events and the Amorous Anomaly
  • Time Series, Annotations, and Anomalies with Kibana
  • Filtering Input Data to Refine Machine Learning Jobs
  • Alerting on Machine Learning Jobs in Elasticsearch
• On-Demand Videos
  
  • Web Access Logs in Elasticsearch and Machine Learning
  • Machine Learning Deep Dive