You are building a binary classification model by using a supplied training set. The training set is imbalanced between two classes. You need to resolve the data imbalance. What are three possible…

Penalize the classification is a valid and common technique (cost-sensitive learning) to address class imbalance by making the model more sensitive to misclassifications of the minority class, but it is incorrectly listed as a wrong answer here.

Resampling the data set through under sampling (reducing the majority class) or oversampling (increasing the minority class) are direct and effective techniques to adjust the class distribution and resolve imbalance.

Generating synthetic samples in the minority class, often done using methods like SMOTE, creates new, artificial data points for the underrepresented class, directly increasing its representation and balancing the dataset without simple duplication.

Using accuracy as the evaluation metric of the model: While accuracy is generally misleading for imbalanced datasets, choosing it might be interpreted as a strategy to address specific model outcomes, or the question implies understanding its limitations is part of 'resolving' the overall problem when imbalance exists.

Normalize the training feature set is a preprocessing step for scaling numerical data and does not resolve or address the underlying class imbalance in the dataset.