Valliappa Lakshmanan

“The Cascade design pattern addresses situations where a machine learning problem can be profitably broken into a series (or cascade) of ML problems. The Ensemble design pattern solves a problem by training multiple models and aggregating their responses. The Neutral Class design pattern looks at how to handle situations where experts disagree. The Rebalancing design pattern recommends approaches to handle highly skewed or imbalanced data.”
― Valliappa Lakshmanan, Machine Learning Design Patterns: Solutions to Common Challenges in Data Preparation, Model Building, and MLOps