Unstructured data analytics fights fraud in California
According to Government Technology, the Los Angeles County Department of Public Social Services has employed new fraud-fighting analytics models based on its vast reserves of unstructured data. The pilot program introducing the technique – which ran between 2008 and 2009 – saved the county $6.8 million.
The source stated that the DPSS' custom platform takes information from various parts of the case management system to create a probability model that can determine if a member of the state's child care system is perpetrating fraud. While the project was in its testing stages, it detected problems with an 85 percent success rate.
Department tech head Michael Sylvester told the source that adoption of the system caused much enthusiasm, with staffers eagerly following the enormous increase in data quality. Integration head Manuel Moreno told the source that the staff found predictive analytics intuitive.
Companies such as eBay already have predictive modeling in place to counter fraud, according to the Wall Street Journal. The web auction giant monitors real-time data based on user insights, including unstructured information, to determine patterns that indicate a user is capable of committing fraud. EBay's system is based on the Hadoop infrastructure, allowing structured and unstructured data to co-exist.