Classification and prediction have numerous applications, including fraud detection.

The outliers may be of particular interest, such as in the case of fraud detection, where outliers may indicate fraudulent activity.

• Example of Data Mining: Fraud detection of credit card usage

1. Credit card companies will alert you when they think your credit card is fraudulently used by someone other than you.

2. Companies will have a history of the customer’s purchases and know geographically where the purchases have been made.

3. If a purchase is made in a city far away from where you live, the companies will put an alert to possible fraud since their Data Mining shows that you don’t normally make purchases in that city.

4. Companies can either disable the card for that transaction or put a flag for suspicious activity.

• Fraud detection detects fraud in applications like Healthcare, retail, credit card service, telecommunications.

• For example :

1. Auto insurance : ring of collisions

2. Medical Insurance : Professional patients, ring of doctors, ring of references.

• Telecommunications: phone call fraud  Analyse the pattern that deviate from an expected norm.

• Retail industry : Analysts estimate that 38% of retail shrink is due to dishonest employees.

• Anti-terrorism

• To detect financial crimes, it is important to integrate information from multiple databases, as long as they are potentially related to the study.

• Multiple data analysis tools can then be used to detect unusual patterns, such as large amounts of cash flow at certain periods, by certain group of customers.

• Useful tools include data visualization tools, linkage analysis tools (to identify links among different customers) Approaches : Clustering, model construction , outlier analysis.