13. Operationalize Machine Learning
• Reinforcement learning is key
• Reduce false positives:
– Recommend incidents to analyst
– Record their workflow
– Build models to learn
• Reward analysts for good analysis:
– Give bonuses for successful investigations
& detailed investigative reports
• Reward machines for good learning:
– Calibrate data, risk/rewards, analyst outcome
19. ML Toolkit & Showcase App
• 20+ standard algorithms for model creation:
– Supervised: Logistic & Linear Regr., SVM, Random Forest
– Unsupervised: KMeans, DBSCAN, Spectral Clustering
• Leverages Python for Scientific Computing
Library and enables use of 300+ algorithms
• Modeling Assistants: Guide model building
and validation
• Showcases: Interactive examples for 20 IT,
security, business, IoT use cases
Extends Splunk platform functions and provides a guided modeling environment
21. Links and Resources
• Download the Splunk App Machine Learning Toolkit and Showcase:
https://splunkbase.splunk.com/app/2890/
• Get started with the latest documentation:
http://docs.splunk.com/Documentation/MLApp/latest/User/About
• Want to dive deeper? Advanced Splunk education course “Splunk for Analytics and Data Science”:
https://www.splunk.com/view/SP-CAAAPCM
• Learn from your peers: 15+ machine learning related sessions at .conf
http://conf.splunk.com/sessions/2016-sessions.html
Lots of helpful resources at your finger tips:
Sept 26-29, 2016 | Walt Disney World Swan and Dolphin Resorts