Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
11
Hunting criminals with hybrid analytics, semi-
supervised learning & agent feedback
David Talby
SVP Engineering
Atigeo
...
2
Why Semi-Supervised Learning & Feedback?
50+Schemes
(and counting)
99.9999%‘Good’ messages
6+Months
per case
3
Why Hybrid Analytics?
Ignore
more rules
Unusual
timing of
eventsUnusual
personal
network
Teamwork
& scale
Think & talk
d...
4
Why Hybrid Analytics?
Rules &
Policies
Time
Series
Analysis
Link
Analysis
Ensemble
Learning
Natural
Language
5
Stream processing
Kafka
Email Stream
Account transactions
Stream
Email NLP
Features
People graph
Transactions time series
6
User Analysis Iteration
Email NLP
Features
User graph
Transactions
time series
Graph Features
Time Series
Features
NLP F...
77
Thank You!
Notebooks for this talk are freely available
David.Talby@atigeo.com
Claudiu.Branzan@atigeo.com
Try xPatterns...
© 2015 Atigeo, Corporation. All rights reserved. Atigeo and the xPatterns logo are trademarks of Atigeo. The information h...
9
10
11
12
13
14
15
16
17
Upcoming SlideShare
Loading in …5
×

Hunting criminals with hybrid analytics strata hadoop v4

495 views

Published on

Presentation on building a machine learning pipeline for hunting criminals using Spark streaming for processing input data and extract message/transaction level features and Python based open source libraries to extract user level features from time series, graphs and unstructured data and use them to train a classifier against agent feedback.

Published in: Data & Analytics
  • Be the first to comment

  • Be the first to like this

Hunting criminals with hybrid analytics strata hadoop v4

  1. 1. 11 Hunting criminals with hybrid analytics, semi- supervised learning & agent feedback David Talby SVP Engineering Atigeo Claudiu Branzan Senior Engineering Lead Atigeo
  2. 2. 2 Why Semi-Supervised Learning & Feedback? 50+Schemes (and counting) 99.9999%‘Good’ messages 6+Months per case
  3. 3. 3 Why Hybrid Analytics? Ignore more rules Unusual timing of eventsUnusual personal network Teamwork & scale Think & talk differently
  4. 4. 4 Why Hybrid Analytics? Rules & Policies Time Series Analysis Link Analysis Ensemble Learning Natural Language
  5. 5. 5 Stream processing Kafka Email Stream Account transactions Stream Email NLP Features People graph Transactions time series
  6. 6. 6 User Analysis Iteration Email NLP Features User graph Transactions time series Graph Features Time Series Features NLP Features Agent Feedback Train/TestClassifier
  7. 7. 77 Thank You! Notebooks for this talk are freely available David.Talby@atigeo.com Claudiu.Branzan@atigeo.com Try xPatterns Connect at: http://xpatterns.com/connect/
  8. 8. © 2015 Atigeo, Corporation. All rights reserved. Atigeo and the xPatterns logo are trademarks of Atigeo. The information herein is for informational purposes only and represents the current view of Atigeo as of the date of this presentation. Because Atigeo must respond to changing market conditions, it should not be interpreted to be a commitment on the part of Atigeo, and Atigeo cannot guarantee the accuracy of any information provided after the date of this presentation. ATIGEO MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
  9. 9. 9
  10. 10. 10
  11. 11. 11
  12. 12. 12
  13. 13. 13
  14. 14. 14
  15. 15. 15
  16. 16. 16
  17. 17. 17

×