Identity Fraud Protection and Big Data Analytics at StampedeCon 2015

August 3, 2015
By Dr. Stephen Coggeshall

Last month, I had the pleasure of attending and speaking at StampedeCon in St. Louis. StampedeCon is a national Big Data conference providing education for both novice IT professionals and big data veterans. My presentation on “Identity Fraud Protection Using Big Data Analytics” outlined the three types of identity fraud that ID Analytics typically sees: 1) Identity Theft, 2) Identity Manipulation and 3) Synthetic Identity. Identity Fraud is the act of misrepresenting which person you are, generally to improperly obtain goods and services or to remain ‘unidentified’ while committing criminal acts such as money laundering. The third type was described in-depth in our 2014 study ‘The Long Con: An Analysis of Synthetic Identities’. Our study showed that while the number of synthetic identities is decreasing, the average fraud rate for synthetic identities has increased more than 100 percent since 2010.

As technology evolves and consumers engage in more transactions online, the greater the potential for fraud.  As individuals use their identities to apply for goods and services – credit, loans, wireless phones, mortgages, etc. – certain patterns of behavior emerge. At ID Analytics, we evaluate these patterns to better understand identity risk.

As part of my presentation, I demonstrated how the different modes of identity fraud are classified by the ID Network and provided examples of fraud rings we have observed. In one example, we uncovered a fraud ring in Indianapolis, IN involving four individuals. Through identity manipulation, these individuals were able to complete 345 applications over a two-year period using 23 different social security numbers and 13 different dates of birth. ID Analytics’ ID Score solution was able to catch 97% of these fraudulent applications. Discovering identity fraud activity in real time requires very careful system architecture and our team utilizes a variety of special technical approaches. We employ a system architecture with separate regions: one optimized for offline data ingestion and preparation, model building and analysis, and the other optimized for low latency for real-time fraud scoring.

Overall, it was a great experience to share what ID Analytics is doing with big data analytics and identity fraud protection with other professionals in the space.

To learn more about the ID Network and our ongoing analysis of fraud rings, download our “I See Fraud Rings” white paper.

Dr. Stephen Coggeshall is Chief Analytics and Science Officer at ID Analytics.