Skip to content

CFA Level I · CFA Level I Exam · Introduction to Financial Data Science

A fund tests a model on a data set in which only 2% of transactions are fraudulent. The model classifies every transaction as legitimate and reports 98% accuracy. Which conclusion is most accurate?

Accuracy is misleading because recall for fraud is 0%. The model never identifies a fraudulent transaction, so it catches none of the actual fraud cases. The 98% accuracy simply reflects that 98% of transactions are legitimate, a classic effect of class imbalance.

  1. ARecall for fraud is 98%, so the model is effective
  2. BAccuracy is misleading because recall for fraud is 0%Correct
  3. CPrecision for fraud is 98% because few false positives occur

Explanation

The model flags no fraud, so true positives are zero and recall for the fraud class is 0/(0+FN) = 0%. Accuracy of 98% merely reflects the class imbalance. Precision is undefined or zero with no positive predictions, not 98%.

Did you get it right without looking?

One question tells you little. A timed set on Introduction to Financial Data Science shows your real accuracy, how long you take and where you lose marks.

Practise Introduction to Financial Data ScienceFree diagnostic mock

More Introduction to Financial Data Science questions