Transcript
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 1/21
Copyright © 2006, SAS Institute Inc. All rights reserved.
Fraud and Risk Management:Universal Responsibility
Andrew Pease
29 April 2008
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 2/21
Copyright © 2006, SAS Institute Inc. All rights reserved.
Agenda
Fraud: Universal Problem/Universal Approach
Approach: Data and Awareness
Rules-Based Alerts
Advanced Analytics
Experiences
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 3/21
Copyright © 2006, SAS Institute Inc. All rights reserved.
Fraud–OneWorld?
What is Fraud? WAKE UP!!! Quiz time!
Ken LayChairman/CEO Enron
Sold 1.8 mil lion shares for more than $101.3million
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 4/21
Copyright © 2006, SAS Institute Inc. All rights reserved.
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 5/21
Copyright © 2006, SAS Institute Inc. All rights reserved.
Finding Fraud is Like……finding a hay-coloured needle in a
haystack…
…scattered all over thefield…
…that doesn’t want tobe found…
…and may not even bethere!
Mission Impossible???
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 6/21Copyright © 2006, SAS Institute Inc. All rights reserved.
Agenda
Fraud: Universal Problem/Universal Approach Approach: Data and Awareness
Rules-Based Alerts
Advanced Analytics
Experiences
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 7/21Copyright © 2006, SAS Institute Inc. All rights reserved.
the iterative process of intelligently filtering large amounts of disparate, yet relevant
data to uncover previously unknown patterns for highly focused investigation
What are ‘Fraud Analytics’?
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 8/21Copyright © 2006, SAS Institute Inc. All rights reserved.
Finding Fraud is Like……finding a hay-coloured needle in a
haystack…
…scattered all over thefield…
…that doesn’t want tobe found…
…and may not even bethere!
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 9/21Copyright © 2006, SAS Institute Inc. All rights reserved.
Agenda
Fraud: Universal Problem/Universal Approach Solution: Data and Awareness
Rules-Based Approach
Advanced Analytics
Experiences
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 10/21Copyright © 2006, SAS Institute Inc. All rights reserved.
Rules-Based Approach
IF-THEN logic
Centralise 80% of FieldBest Practices
Investigation Framework
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 11/21Copyright © 2006, SAS Institute Inc. All rights reserved.
Discovery
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 12/21Copyright © 2006, SAS Institute Inc. All rights reserved.
Agenda
Fraud: Universal Problem/Universal Approach Solution: Computers and Awareness
Rules-Based Approach
Advanced Analytics
Experiences
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 13/21Copyright © 2006, SAS Institute Inc. All rights reserved.
Subsidized Amount
I n c
o m e
cluster4
cluster3
cluster1
cluster2
cluster5
Segmentation/Clustering
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 14/21Copyright © 2006, SAS Institute Inc. All rights reserved.
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 15/21Copyright © 2006, SAS Institute Inc. All rights reserved.
Decision Tree
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 16/21Copyright © 2006, SAS Institute Inc. All rights reserved.
SAS Fraud Detection
Differentiating Between Monitoring Approaches
Rules
Set up rulesto filter suspicious
transactions
Examples:Refund > 20.000
AND
Income < 30.000
Profiling
Build statisticalProfiles of
Entities andinteractions
Examples:Mean, standard
deviation, quartiles,
distributions
AdvancedAnalytics
Knowledge discovery
In databasesand machine
learning
Examples:Neural networks
Fuzzy logicGenetic algorithms
Hybrid
Combination ofall existing
approaches
Examples:Genetic algorithm and
Statistics plus
Neural network
Issues
• Detect known andunknown patterns offraudulent behavior
• Keep track with newpatterns Not exactlyknowing what to lookfor
Suitable forknown
patterns
Suitable forunknown
patterns
Suitable forcomplex
patterns
Best
Practice
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 17/21Copyright © 2006, SAS Institute Inc. All rights reserved.
Agenda
Fraud: Universal Problem/Universal Approach Solution: Computers and Awareness
Rules-Based Approach
Advanced Analytics
Experiences
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 18/21Copyright © 2006, SAS Institute Inc. All rights reserved.
KBC Internal Fraud
INPUTData
THROUGHPUTAnalysis & mining
OUTPUTApplication
AFK
TMK
...
Known rules
KRD
New rules
Intelligence
Server
SAS EM
KBC LOA fraud risk
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 19/21Copyright © 2006, SAS Institute Inc. All rights reserved.
Fraud Investigation at OLAF
ACCESS Products
Enterprise Miner
(incl. Text Miner)
ETL Business Intelligence Analytic Intelligence
SAS Data Set s
Cluster ing
Classi f icat ion
HTML
OCR
Reports
Tex t
Preprocessing
MS Word Oracle
Registration Management
Improved Case Follow-up
Access
Enterprise Guide
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 20/21Copyright © 2006, SAS Institute Inc. All rights reserved.
Banksys Fraud Detection Approach
Data
Capture
OperationalSystems
Data
Transfer
FD&RServer
CMT
PC
SAS Fraud Detection
Engine
Data
Upload
FD&R Database
Alerts and InvestigationInformation
CardStop, ...
8/6/2019 6 Pease Andrew Dealing With Large Data
http://slidepdf.com/reader/full/6-pease-andrew-dealing-with-large-data 21/21
AFK
TMK
...
Known rules
KRD
New rules
Intelligence
Server
SAS EM
KBC LOA fraud risk
D a t a
C a p t u r e
O p e r a t i o n a lS y s t e m s
D a t a
T r a n s f e r
F D & RS e r v e r
C M T
P C
S A S F r a u d D e t e c t i o nE n g i n e
D a t a
U p l o a d
F D & R D a t a b a s e
A l e r t s a n d I n v e s t i g a t i o nI n f o r m a t i o n
C a r d S t o p , . . .
top related