You are using a machine learning algorithm for $1$st round of cancer detection.

For this algorithm $-$ False Positive $($classifying a fit $($normal$)$ person as having cancer$)$ is more expensive than False Negative $($Classifying a patient having cancer as fit$)$