#SyntheticIdentityFraud costs the financial industry approximately $1.5B per year and is growing. This criminal methodology happens when someone uses a combination of real and fake personal information to create an identity and commit fraud, and by criminals to hide their identities from discovery in many market segments to include the use by foreign intelligence services and international criminal organizations. How can U.S. commercial and government sectors discover these synthetic identities, fraud, and other nefarious activities?
Advanced Onion Inc. (AO) uses a combination of methods and techniques for different data types and sources based on the type of discovery needed. For single data sets, simple techniques using automated exploratory data analysis with statistical methods like standard deviations, Bedford's law, frequency distributions, and trend analysis may suffice. For multiple datasets, #K-Means, #DBSCAN, #IsolationForest, and #Autoencoders algorithms, may be used for unsupervised data discovery. A critical issue is the necessity for “right-to-explanation” proofs that measure the contributions of input variables to an ultimate algorithmic output.
AO has been involved in the people analytics business since 2011, with our DoD Mirador #ContinuousEvaluation system, ongoing #Biometrics support to law enforcement agencies, and numerous other identity-related programs. As we continue to evolve our services supporting people analytics and discovery, we are developing leading edge capabilities to objectively and repeatably evaluate these concepts.
Comments