Retraining
Domain: MLOps
🎯 Motivation
Training the model again with new data. Essential when drift is detected or new data is available.
📋 Concrete Example
Retraining: monthly. New data from the past month is added. Model retrains. New version: v4 (Jul, 92%). Deployed. Performance maintained.
🔗 Analogy
A student reviewing new material each month - they learn new things and update their knowledge. Retraining updates the model with new data.