There are several things to consider when evaluating an ML-based fraud detection model. These factors include cost, effectiveness, and reliability. Amex has recently announced a new program to fight fraud with its ML-based model. It’s early days to tell how successful this program will be in combating fraud, but Amex says the new program will help it stay a leader in this field. Stay tuned for more information.
ML-based fraud detection model
The new ML-based fraud detection model developed by American Express is meant to be more effective than the current rule-based approach. The new model was created using billions of observations and a series of 1,000 decision trees. It can make decisions in milliseconds, reducing the chance of making a mistake while still maintaining the highest quality of service. The new system is already a big step forward for Amex, which has long been one of the lowest-fraud rates in the credit card industry.
The Gen X model is the largest model used by American Express to monitor fraud risk. It was first iterated in 2014 and has grown to be one of the biggest models used by credit card company. To build it, AmEx began by prioritizing a business question. After identifying a critical business question, the team built the model. It evaluates more than 8 billion transactions a year, considering the customer’s account history and past payment history.
Despite the growing popularity of the model, Amex has not been able to deploy it at full capacity. The company’s cloud provider, Amazon Web Services, did not appear to have enough capacity to handle the influx. Amex has not yet released a number of usage figures. In the meantime, it continues to tweak its model based on human training. This article explores the process by which the new ML-based fraud detection model came to be used at American Express.
A new database startup has helped Amex transfer its ML-based fraud detection model from its predecessors. Compared to other financial services companies, AmEx has fewer “silos” than other institutions. This makes it easier to transfer machine-learning models across to the new company. The data used for building the model should be real-time and not stored in the past. The model should be able to identify fraudulent behavior quickly and accurately.
Competitiveness
Amex has been pursuing ML-based solutions for years, and is now introducing one of them to its fraud detection system. Amex is now attempting to improve the fraud detection process by using a new model that is designed to be more efficient than the rule-based system that is currently in place. As more financial institutions adopt ML-based solutions, Amex’s business is increasingly under pressure to keep up.