Interview Guides

You're training a supervised classification model, but the positive class is much rarer than the negative class. You want the model to detect the minority class reliably without being misled by accuracy.
How do you address class imbalance in a dataset when training a predictive model?