Optimize a machine learning model for imbalanced data

You are given a dataset with a significant class imbalance (e.g., 95% of the data belongs to class A, and 5% belongs to class B). Design a strategy to train a classification model that performs well on both classes. Discuss techniques to handle the imbalance, such as sampling methods, cost-sensitive learning, or evaluation metrics that should be prioritized. Assume the dataset contains 10,000 samples, and the model will be evaluated on a separate test set.