Argyle Data uses state-of-the-art machine learning on a Hadoop stack reducing the window for risk or fraud from hours to minutes
Traditional analytics approaches are batch oriented bringing together a patchwork quilt of database technologies, ETL, and BI tools. Most risk applications based on a previous generation stack, come with a high cost database infrastructure. This patchwork quilt often necessitates in a long wait for nightly ETL and transformation jobs to be completed to give insights that are hours or days out of date. Headquartered in San Mateo, CA, Argyle Data uses state-of-the-art machine learning on a Hadoop stack to deliver risk applications that can ingest data and analyze it in real time, reducing the window for risk or fraud from hours to minutes.
Working with some of the biggest companies in the world from the fields of mobile communications, e-commerce, and finance, Argyle Data’s simple queries and machine learning at petabyte scale can guide a user to identify existing fraud and new fraud as it is emerging. The key factor in fighting risk in the mobile space is in reducing the time window risk coefficient. In 2013 the mobile communications industry saw a 15 percent increase in fraud from that of 2011, losing $46.3 billion in the process. Fraud occurs in many forms–Wangiri, SMS phishing, roaming, premium rate service, international revenue share, to name a few. For Argyle Data customers, the rationale behind the savings of more than $1 million per month in a typical operating company or country is the ability in detecting all kinds of frauds in real time as opposed to a day later.
“Risk and fraud is strategic to security on a global scale. We intend to take our unique approach to machine learning risk applications on a Big Data stack to additional customers in the U.S. and Europe and plan to continue our expansion throughout these regions and into Asia over the next few years,” adds Ryan.