State of the Art Technology Delivers Better Results
To ensure ID Analytics’ products and solutions exceed our customers’ expectations, ID:A Labs continuously conducts technical research on new algorithms, variables, methodologies, data sets and business applications. Here are a few examples of areas we’re currently exploring.
The ability to see data and its relationships aids greatly in discovery of behaviors and trends – building our understanding of U.S. identity dynamics and revealing opportunities for product development. The continuous exploration of new methodologies and technologies enables ID:A Labs to better visualize complex data relationships. Data visualization can be used to better understand the complex relationships in credit risk, the existence of the underlying identity, and first party authenticity which we derive from the ID Network®.
Our data is complex, with ID3D we’re able to observe subtle behavioral differences which reveal more types of fraud such as hidden synthetic identities that are not typically detected with standard review methods. Using this approach we are able to gain a holistic view of identity risk.
3D identity profiling provides a holistic view of the incoming identity and the application. It can help our clients to make more intelligent decisions by providing unique insights into the nature of their population’s risk. This platform could also help our clients monitor fraud movement across different industries (banking, retail, wireless), channels (offline, inline or mobile), and modes of bad behaviors. By periodically re-clustering the solution space we work with our clients to discover new clusters or new types of bad behaviors, to improve policies for remediation.
Deep Neural Nets
With ever-increasing hardware capabilities, deep learning using neural networks has become an important resurgent technology. Initially inspired by an understanding of the deep architecture of cortical visions systems, deep learning allows supervised formation of internal structures that can automatically learn features in data at increasing levels of abstraction as the depth of the architecture increases. This capability allows ID Analytics scientists to leverage the rich structure of the ID Network® — building on an already rich set of features that number in the tens of thousands. This allows ID Analytics to build on already best in class performance.
Scalable and Hierarchical Identity Resolution
The ID Network® encompasses billions of identity-based records. Each of these records represents information which conveys aspects of an identity. Best representations of an identity are derived by collectively integrating all the records that pertain to that identity. Given the immensity of the ID Network®, determining that set of records is an intractable problem that would require quintillions of comparisons using brute force methods. Using proprietary technologies, ID Analytics is able to reduce the search space by many orders of magnitude to find the detailed set of records which are relevant. We use hierarchical clustering methods to discern identities and higher level constructs smoothly and accurately move across hierarchical entity levels, going from individual events to accounts, people and households. All this is achieved using massively parallel graph processing systems.
Value of Hi-Fidelity Identity Resolution for Credit Risk Measurement
In predictive analytics and risk analysis data is king. The most complex predictive systems are always dependent on the amount and fidelity of information that they receive. An understanding of risk benefits from as complete a picture of an identity as possible. ID Analytics’ ability to leverage the full scale of the ID Network® benefits from its core identity resolution capabilities.