What does data drift refer to in data science?

Enhance your skills with the CompTIA Data+ Certification Test. Engage with flashcards, tackle challenging multiple choice questions, complete with hints and explanations. Get yourself exam-ready now!

Data drift refers to a change in the statistical properties of the data the model encounters after deployment. When the distribution of input features shifts over time, or the relationship between features and the target changes, the model’s predictions can become less accurate because it was trained on a different data distribution. This is different from other issues: model bias is about systematic errors in predictions due to the model itself, overfitting is when a model captures noise from the training data and generalizes poorly, and data normalization is simply a preprocessing step to scale features. Understanding data drift helps you monitor and update models as the real-world data environment evolves, including cases like covariate drift (changes in feature distributions) and concept drift (changes in the feature-to-target relationship).