Amazon Fraud Detector now supports Batch Fraud Predictions
Today, we are happy to announce that Amazon Fraud Detector now supports batch fraud predictions. Amazon Fraud Detector is a fully managed service that makes it easy to identify potentially fraudulent online activities, such as the creation of fake accounts or online payment fraud. Until today, Fraud Detector supported real-time fraud predictions via the GetEventPrediction API, which is ideal for low-latency synchronous fraud prediction use cases but requires customers to integrate an API and make an API call for every event they want to evaluate. Now, customers who have non real-time fraud prediction use cases can get fraud predictions for a large number of events in one go, or on an hourly or daily basis, using the new batch prediction feature without needing to write any code.
Today, we are happy to announce that Amazon Fraud Detector now supports batch fraud predictions. Amazon Fraud Detector is a fully managed service that makes it easy to identify potentially fraudulent online activities, such as the creation of fake accounts or online payment fraud. Until today, Fraud Detector supported real-time fraud predictions via the GetEventPrediction API, which is ideal for low-latency synchronous fraud prediction use cases but requires customers to integrate an API and make an API call for every event they want to evaluate. Now, customers who have non real-time fraud prediction use cases can get fraud predictions for a large number of events in one go, or on an hourly or daily basis, using the new batch prediction feature without needing to write any code.
Batch prediction jobs calculate fraud predictions based on the customer’s configured detection logic, in the same way as the GetEventPrediction API. With a batch prediction job, you point Fraud Detector to an input file stored in S3 that contains the events you want to evaluate. The batch prediction job reads your input file, generates fraud predictions for each event, and outputs the results, including model scores and rule outcomes, to an S3 location of your choice. Each batch prediction job can support input files up to 50 megabytes (MB), or about 50K events (varies by event size and detection logic). You can use the batch prediction feature via the AWS console, SDK or CLI.
To create a batch predictions job using the console, sign in to the AWS Management Console, open the Amazon Fraud Detector console, and navigate to ‘Batch predictions’ on the left navigation menu. Next, choose New batch prediction and then configure the job name, the Detector you want to use, the IAM role with the required S3 permission, and the input and output file location. Once complete, click Start. When the job is finished, the output file will appear in the specified S3 location.
Fraud prediction pricing is the same whether customers choose to use the real-time GetEventPrediction API or the new batch prediction feature. For more information on Fraud Detector pricing, visit our pricing page. The new batch prediction feature for Amazon Fraud Detector is available today in the US East (N. Virginia), US East (Ohio), US West (Oregon), Europe (Ireland), Asia Pacific (Singapore) and Asia Pacific (Sydney) regions. For further details, see our documentation.