Causal view reveals different causes behind concept drift in data
Concept Drift from a Causal Perspective
Machine Learning
Summary
Predictive models often struggle when the data they see changes over time, a problem called concept drift. The authors look at this problem through the lens of cause and effect, identifying different reasons why the data might shift. They build a tool to create data with specific types of drifts based on these causes and show that each type changes model predictions differently. This approach helps make testing adaptive models more realistic by reflecting how real-world data changes.
What this means in practice
- •For machine learning engineers: Generate realistic data streams with different drift causes to test and improve adaptive models.
- •For data platform developers: Create synthetic data that mimics real-world dependency changes to better evaluate data stream handling systems.
Authors
Eduardo V. L. Barboza, Jean Paul Barddal, Robert Sabourin, Rafael M. O. Cruz
Abstract
Concept drift is a common phenomenon in real-world data streams, in which changes in the data-generating distribution can degrade predictive model performance. Most existing definitions characterize drift as changes in the joint distribution $P(\mathbf{x}, y)$, without distinguishing which component of the data-generating process has changed. In this work, we introduce a causal perspective on concept drift based on Structural Causal Models (SCMs). We propose a taxonomy that categorizes drift events by their causal origin, including changes in exogenous variables, endogenous mechanisms, confounders, and target-generating processes. Building on this framework, we develop an SCM-based data stream generator that simulates controlled mechanism-level drift events. Our experiments empirically characterize the distributional effects of each drift type and show that drifts with different causal origins induce distinct patterns of distribution shift and predictive behavior. Furthermore, by integrating causal discovery methods, we use our framework to construct data streams grounded in real-world dependency structures, enabling more realistic and informative evaluation scenarios. We also demonstrate that leveraging the generated data can improve downstream performance. These results highlight the importance of accounting for causal structure when studying and evaluating adaptive learning methods, and establish a foundation for causally-aware evaluation in non-stationary environments.