Microsoft announced today that the accuracy of its Azure Active Directory Identity Protection detection algorithms has been increased by 100% and the false-positive rate has been decreased by roughly 30%. Microsoft’s identity and access management solution is designed to be used with on-premises, cloud-based and hybrid implementations. The new and improved Unfamiliar Sign-in Properties triggers three risk levels, from no risk to high risk, each of them associated with the likelihood of the analyzed authentication being compromised based on a risk score.
Source: https://www.bleepingcomputer.com/news/microsoft/microsoft-boosts-compromised-account-detection-in-azure-ad-by-100-percent/