Fraud & Anomaly Detection using Data Mining To Manage Risk

Data Mining is the use of statistical and logical techniques to “uncover hidden patterns and relationships” within large sets of structured or unstructured data records and to “infer decision rules” useful in predicting future outcomes.

Delivery Expectations

Reports and intelligence which expressly uncover fraud / trails and report anomalies which enable business analysts and / or forensic auditors to clarify the full facts for prevention and recovery.

It has been identified that fraud is perpetrated by three main groups within companies:

• Internal staff on their own
• External parties on their own
• Collusion between Internal and External parties

Statistics:

“42% of companies experience fraud based upon poor data quality & validation” – (Survey by “Kroll”)

“30% of all fraud is found via analysis vs. tips” (Coderre, “Fraud Detection”)

Processes

• Identifying errors and potentially fraudulent patterns or trends by EI data analysts which would otherwise be difficult to detect in traditional paper-based investigations
• Comparing patterns/trends
• Serving as a possible fraud deterrent
• Assisting in data clean-ups and providing best practices/services to keep data clean

Effective Intelligence has developed a software suite, InfoArchitect™, which standardises, checks and corrects all data elements within consumer datasets.

Sources for Deceased

Clients who have traditionally used the Home Affairs database have found that matching to Effective Intelligence is a more effective method as the source dates are far more recent, which otherwise would be captured many months after the person has deceased.

Fraud is often committed by obtaining the ID’s, Title Initial, Surname and Addresses from legal notices. Recent studies have proved a reduction in fraud for those companies using EI solutions, by using multiple verification sources. Along with InfoArchitect™ validation software suite, EI is able to identity duplicate data fields and flag these at a much earlier stage.

Data Interrogation Concepts

• Master Files – changes, duplications, missing, illogical issues
• Pattern, Frequency, Range – identifying anomalies
• Mistakes – basic mistakes indicating fraud
• Duplicate Analysis – identify genuine duplication
• Illogical – frequency and patterns that are not normal
• Trends – abnormal increasing or decreasing trends
• “Change” patterns – new, delete, update, void etc.
• Circumvention Strategies – processing below control thresholds
• Covert vs. Overt – deliberate changes designed to prevent automated detection
• Data Interpretation – experienced review of data looking for new / unusual patterns

Example Fraud Indicators

• Ghost employees but matching deceased individuals (EI Deceased)
• Duplicate employees at similar contact details (IA Address Matching, etc.)

Examples of Ghost Employees

a) Terminated ghost employees
b) Fictitious / deceased ghost employees
c) No-show ghost employees
d) Temporary ghost employees
e) Family ghost employees
f) Rehired ghost employees
g) Pre-employment ghost employees

EI & External Data Sources

Internal/External Data:
• Deceased persons
• Spatial Household / Families – match for relationships

Deceased identity fraud:
• Fraud Scheme – Register temporary or permanent “workers” using deceased identities and “Resign” them before data made available by DTI

Staff originated vendor fraud:
• Fraud Scheme – Staff uses altered contact details (similar to employee details) for false vendor entity that then bills and gets paid • EI solves this by re-structuring, validating and converting all contact details to structured English and then originating ID’s at various levels to dramatically improve matching routines (InfoArchitect™)

EI Strengths

• Operated since 1998
• Trusted track record of dealing with major companies holding consumer databases
• Data bureau processes over 1 billion records per annum
• Developed leading proprietary InfoArchitect ™software to fix data and derive matching address and contact detail ID’s
• Data comparison and matching algorithms built over 15 years
• Full-time data bureau staffed with skilled SQL technicians
• Full-time internal development team building EI software
• Experienced Business Analysts and Master Statisticians
• Major data assets include a significant Consumer Household Database with unique attributes
• Supplier of specialised data software to Credit Bureaus, Financial Service Providers, Insurance companies, Retailers, etc.
• Accredited by DMA SA as a “Database Centre of Excellence”
• Accredited by CIMA as a “CIMA Training Partner”
• Level 3 BBEEE certification • Full SARS Tax Clearance certification
• Non-executive Directors – Ms. Christiane Duval (Legal data privacy expert) – Mr. P Baines (CA and data usage expert)