How do you calculate the False Alarm Rate (False Positive Rate) in a binary classification scenario?