Understanding and detecting different types of drift | by Ajay Verma | Medium

In machine learning, “drift” refers to changes in the data distribution over time, which can affect the performance of models. Understanding and detecting different types of drift is crucial for maintaining model accuracy and reliability. Here are the primary types of drift in machine learning: Concept drift occurs when the relationship between the input features (X) and the target variable (y) changes over time. This type of drift affects the predictive model’s performance because the underlying patterns it has learned no longer hold true. Concept drift can be further categorized into: Sudden Drift: The relationship changes abruptly. Example: A sudden change in customer behavior due to a new competitor entering the market. Gradual Drift: The relationship changes slowly over time. Example: Slowly evolving user preferences on a social media platform. Incremental Drift: The relationship changes in small, incremental steps. Example: Gradual changes in sensor readings due to wear and tear

saved by

• Enrique Moran