DataVisor, and AI eCommerce fraud prevention solution, announced the launched of its new Feature Platform in a press release. The platform enhances merchants ability to use data and analytics to create features to prevent eCommerce fraud.
“The DataVisor GIN is powered by derived signals from more than 4.2 billion protected accounts and in excess of 800 billion events across industries,” the press release said. “This significantly enhances machine learning with fine-grained digital intelligence that encompass everything from IP address patterns to user agent strings and more.”
There are two aspects of the platform. One that uses internal data, and one that uses known best practices.
The internal data option allows fraud analysts to build new features from internal data quickly. The platform analyzes trends and creates features to combat vulnerabilities. In addition, teams can also benefit from built-in features known to work for specific types of fraud. These features engage automatically based on scenario analysis.
“DataVisor solution can recommend features that are already optimized for specific use cases, particularly to address risk and fraud issues,” the press release said. “For example, if the organization focuses on transaction fraud, Feature Platform will recommend a list of features that are readily available and uniquely important to that scenario to deliver strong detection results immediately.”
In the statement, Yinglian Xie, DataVisor’s CEO, stressed that AI is critical to preventing fraud loss. In addition, he stated that DataVisor takes a holistic approach to the analysis of data underpinning its fraud solution. Overall, he believes the combination means the new Feature Platform will create effective models faster than current eCommerce fraud prevention technologies.
“The hallmark of the DataVisor approach is our holistic approach to data analysis using sophisticated feature engineering and machine learning,” he said. “The Feature Platform incorporates this ethos and delivers organizations with proven means to develop effective machine learning models more easily and faster, and hence stop fraud loss from occurring.”