Best kept machine learning secret in security

The allure of using machine learning in data security comes from its ability to generalize attack detection based on historical data and to detect attacks that would not be obvious otherwise. Machine learning in security analytics is gaining widespread adoption, and the security analytics market is projected to hit 7.1 billion by 2020. The biggest challenge in using machine learning for data security has to do with triaging, or prioritizing, alerts effectively. In my last post, I explored how to prevent false alerts in data security. Here, we'll explore how a generalizable algorithm-based system can detect security breaches, using ranking algorithms from the information retrieval domain.