Economists were the first to tackle the issue of insurance fraud. They designed a set of optimal contracts discouraging claimants to fraud. These contracts require a loss region with a credible probability of audit. Given the volume of claims faced, automated systems used to flag patterns of fraud are a requirement. The recent explosion of new ML expert systems makes us forget that an audit decision is an economic problem. By embedding ML algorithms into a structural economic model which explains the claimant’s decision to fraud, we can get the better of both worlds.
- Күн бұрын
Raphaël Zerbato - AI for better fraud detection
- Рет қаралды 33
Пікірлер