Definition:
Machine learning fraud detection systems use artificial intelligence solutions to detect ‘acts of fraud’. These techniques fall under two main categories:
- Supervised learning systems – these systems require training data sets to learn and use techniques like neural networks, regression models, statistical models, or a combination.
- Unsupervised learning systems – these systems are able to identify potential fraud based on techniques like clustering, peer group analysis, breakpoint analysis, pro ling or a combination.