A
You are working on a supervised classification problem where the positive class is rare, and a standard accuracy score would hide most of the mistakes. The team needs a model that can find the minority class without flooding operations with false alarms.
How would you handle a highly imbalanced dataset?
