Data preparation

Feature engineering is performed during the data preparation phase to transform raw data into informative features that help the model learn patterns during training. Option A is incorrect because feature engineering must occur before training so the model can learn from these engineered features, even though the same transformations are later applied at inference.

Data leakage occurs when target information from the test or validation set inadvertently influences the training process, leading to unrealistically high evaluation metrics that fail in production. This is a machine learning methodology issue rather than a security event like a network breach.

Class imbalance causes the model to over-represent the majority class, leading to poor recall for the minority class and biased, unfair predictions. A diagram for this topic isn't available yet — the worked example below walks the same reasoning step by step.

• !Option A is incorrect because feature engineering must occur before training so the model can learn from these engineered features, even though the same transformations are later applied at inference.

Answer each from memory — you'll see instantly whether you're right and why.

Run a focused 10-question mini-mock on Data preparation and see it stick.