2 DATA MINING FOR FRAUD
In 2002, Gene Morse found a round $500 million debit to a PP&E account at
WorldCom. He discovered the anomaly through searches in a custom data warehouse he
had developed in the Essbase multidimensional database. WorldCom would not give
Morse access to full financial systems, so he created his own warehouse and used basic
data mining techniques to search it. Using a small script and Microsoft Access, Morse
followed the account through the financial reporting system and ultimately discovered a
$1.7 billion entry of capitalized line costs in 2001 (Lamoreaux, 2007).
The WorldCom fraud discovery is one example of using computer technology to
search full populations of data for anomalies, trends, and fraud. Traditional auditing uses
techniques like discovery, stratified, or random sampling to determine whether a
population contains errors (Albrecht and Albrecht, 2002). This approach works well