問題文
A retail bank holds several years of card transaction records. Nobody has ever marked which of those transactions were fraudulent, and the analytics team wants to start by discovering which behaviors look alike so that investigators can inspect the unusual groups first. Which approach fits what the team actually has to work with?
選択肢
- The clustering process that can group the unlabeled transactions, because the labels available in the data settle the choice.
- The supervised predictor that can predict the labeled outcomes, because the labels available in the data settle the choice.
- The recommendation engine that can order the scored listings, because the labels available in the data settle the choice.
- The regression model that can forecast the monetary values, because the labels available in the data settle the choice.