Professional Data EngineerOperationalizing machine learning modelsEasy
A retail company has deployed a personalized marketing ML model. After a few months, the model's recommendations start to become less effective, leading to a decrease in conversion rates. Upon investigation, they discover that customer preferences and purchasing behaviors have significantly changed since the model was last trained, rendering the original learned patterns less relevant. What phenomenon is their model experiencing?
- AOverfitting
- BUnderfitting
- COutlier detection
- DConcept drift
Show answer & explanationAnswer & explanation
Correct answer: D. Concept drift
Concept drift occurs when the relationship between the input features and the target variable changes over time, meaning the underlying 'concept' the model is trying to learn has shifted, which aligns with the scenario of changing customer preferences and purchasing behaviors.
Why the other options are wrong
- A. Overfitting is when a model learns the training data too well, including noise, and performs poorly on unseen data, which typically happens during training, not a gradual degradation in production due to changing underlying concepts.
- B. Underfitting is when a model cannot capture the underlying patterns in the training data, leading to poor performance on both training and test data, usually due to a model that is too simple.
- C. Outlier detection is about identifying anomalous data points, not a degradation in model performance due to changing underlying patterns.
Concept Drift
A phenomenon in machine learning where the statistical properties of the target variable, which the model is trying to predict, change over time, leading to a degradation in model performance.
- The relationship between features and target changes.
- Requires model retraining or adaptation.
- Common in dynamic environments (e.g., customer behavior, financial markets).
Memory trick: When the concept drifts, the model gets lost.