Class Imbalance: Monitoring Rare and Nuanced Model Drift

Many types of machine learning models, from fraud to advertising to eCommerce, involve an imbalance between classes with one occurring far more frequently than another, e.g. instances of no fraud versus fraud. Detecting drift in the minority class can be a difficult but critical business requirement.

Learn what class imbalance is and how to detect it, including:

• Why class imbalance occurs
• The impact of class imbalance on ML models
• How to address class imbalance for effective model monitoring