slides presented in MTI-820 (BI) graduate course (ETS) as an invited speaker

1. Outline
Why a talk about machine learning and BI?
Machine Learning 101
Let’s dive into practical example
Conclusion
Questions?
No BI without Machine Learning
Francis Pieraut
francis@qmining.com
http://fraka6.blogspot.com/
10 March 2011
MTI-820 ETS
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

2. Outline
Why a talk about machine learning and BI?
Machine Learning 101
Let’s dive into practical example
Conclusion
Questions?
To Much Data
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

3. Outline
Why a talk about machine learning and BI?
Machine Learning 101
Let’s dive into practical example
Conclusion
Questions?
Why a talk about machine learning and BI?
Machine Learning 101
Supervised Learning (classification)
Unsupervised Learning (clustering)
Training and Testing
Important Concepts
Let’s dive into practical example
Target Marketing
Customer behavior
Retention
Risk Analysis
Monitoring Root Cause Analysis - QMonitor
Monitoring Root Cause Analysis - QMiner
Conclusion
Questions?
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

4. Outline
Why a talk about machine learning and BI?
Machine Learning 101
Let’s dive into practical example
Conclusion
Questions?
Why a talk about machine learning and BI?
Machine Learning ⇒ Data-Mining ⇒ BI
Prediction or Clutering ⇒ Patterns ⇒ Patterns (revenus $$ ⇑)
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

5. Outline
Why a talk about machine learning and BI?
Machine Learning 101
Let’s dive into practical example
Conclusion
Questions?
Speaker: Francis Pieraut, P.Eng. M.Sc.A.
Master@LISA - Statistical Machine Learning - udm
(flayers: C++ Neural Networks lib)
Industry - 7 years in Machine Learning/AI startups
(mlboost: Python Machine Learning Boost lib)
Founder QMining
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

6. Outline
Why a talk about machine learning and BI? Supervised Learning (classification)
Machine Learning 101 Unsupervised Learning (clustering)
Let’s dive into practical example Training and Testing
Conclusion Important Concepts
Questions?
AI and Machine Learning - Data-mining
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

7. Outline
Why a talk about machine learning and BI? Supervised Learning (classification)
Machine Learning 101 Unsupervised Learning (clustering)
Let’s dive into practical example Training and Testing
Conclusion Important Concepts
Questions?
Machine Learning and Data-Mining
Machine Learning: learn from data
Data-mining: extracting patterns from data
Machine Learning use extracted patterns to do prediction
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

8. Outline
Why a talk about machine learning and BI? Supervised Learning (classification)
Machine Learning 101 Unsupervised Learning (clustering)
Let’s dive into practical example Training and Testing
Conclusion Important Concepts
Questions?
Machine Learning
Learning from data
Classification vs Clustering
Applications: Attrition, Rank Customer (approve loans and
credit card),Fraud detection, Target-Marketing, Risk Analysis
(insurance) etc.
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

9. Outline
Why a talk about machine learning and BI? Supervised Learning (classification)
Machine Learning 101 Unsupervised Learning (clustering)
Let’s dive into practical example Training and Testing
Conclusion Important Concepts
Questions?
Supervised Learning (need class tag for each example)
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

10. Outline
Why a talk about machine learning and BI? Supervised Learning (classification)
Machine Learning 101 Unsupervised Learning (clustering)
Let’s dive into practical example Training and Testing
Conclusion Important Concepts
Questions?
Unsupervised Learning - dimension reduction/clustering
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

11. Outline
Why a talk about machine learning and BI? Supervised Learning (classification)
Machine Learning 101 Unsupervised Learning (clustering)
Let’s dive into practical example Training and Testing
Conclusion Important Concepts
Questions?
Learning Process
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

12. Outline
Why a talk about machine learning and BI? Supervised Learning (classification)
Machine Learning 101 Unsupervised Learning (clustering)
Let’s dive into practical example Training and Testing
Conclusion Important Concepts
Questions?
Tanks in the desert (black box danger)
Using ML requires insights
An algo is only goods as its data
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

13. Outline
Why a talk about machine learning and BI? Supervised Learning (classification)
Machine Learning 101 Unsupervised Learning (clustering)
Let’s dive into practical example Training and Testing
Conclusion Important Concepts
Questions?
Important Concepts
Datasets (features + class)
Generalization vs Overfitting
Classification vs Clustering
Features Quality (invariant and informative)
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

14. Outline Target Marketing
Why a talk about machine learning and BI? Customer behavior
Machine Learning 101 Retention
Let’s dive into practical example Risk Analysis
Conclusion Monitoring Root Cause Analysis - QMonitor
Questions? Monitoring Root Cause Analysis - QMiner
Service provider
Find most probable interested clients
N most likely to buy (sort DESC probability)
google mail
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

15. Outline Target Marketing
Why a talk about machine learning and BI? Customer behavior
Machine Learning 101 Retention
Let’s dive into practical example Risk Analysis
Conclusion Monitoring Root Cause Analysis - QMonitor
Questions? Monitoring Root Cause Analysis - QMiner
Cell phone usage
Find users cluster (behavior)
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

16. Outline Target Marketing
Why a talk about machine learning and BI? Customer behavior
Machine Learning 101 Retention
Let’s dive into practical example Risk Analysis
Conclusion Monitoring Root Cause Analysis - QMonitor
Questions? Monitoring Root Cause Analysis - QMiner
Service provider
Find most probable clients to quit
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

17. Outline Target Marketing
Why a talk about machine learning and BI? Customer behavior
Machine Learning 101 Retention
Let’s dive into practical example Risk Analysis
Conclusion Monitoring Root Cause Analysis - QMonitor
Questions? Monitoring Root Cause Analysis - QMiner
Insurance
Score customer risk of making a claim
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

18. Outline Target Marketing
Why a talk about machine learning and BI? Customer behavior
Machine Learning 101 Retention
Let’s dive into practical example Risk Analysis
Conclusion Monitoring Root Cause Analysis - QMonitor
Questions? Monitoring Root Cause Analysis - QMiner
QMonitor - Global Server Incident Mining
Find incidents on servers
Find patterns (network, server, etc.)
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

19. Outline Target Marketing
Why a talk about machine learning and BI? Customer behavior
Machine Learning 101 Retention
Let’s dive into practical example Risk Analysis
Conclusion Monitoring Root Cause Analysis - QMonitor
Questions? Monitoring Root Cause Analysis - QMiner
QMiner - Global User Experience Incident Mining
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

20. Outline
Why a talk about machine learning and BI?
Machine Learning 101
Let’s dive into practical example
Conclusion
Questions?
What you should remember?
No BI without Machine Learning
Machine learning algorithms applications ⇑
goal = generalization⇒good prediction (DON’T OVERFIT)
80-90% pre or post-processing + data visualization
Python provide amazing integration
**QMining is looking for intership students
BI for Business User http://www.qlikview.com/
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning

21. Outline
Why a talk about machine learning and BI?
Machine Learning 101
Let’s dive into practical example
Conclusion
Questions?
Any questions?
...
intership 2011 ⇒ francis@qmining.com
http://fraka6.blogspot.com/
..
Thanks,
Francis Pieraut
francis@qmining.com
Francis Pieraut francis@qmining.com http://fraka6.blogspot.com/ No BI without Machine Learning