FRM Part I · FRM Exam Part I · Machine Learning and Prediction
A bank's credit-default classifier is tested on 200 loans. It flags 50 loans as defaults, of which 40 actually defaulted. In total 60 loans in the sample defaulted. What is the precision of the classifier?
Precision is 0.80. It is the share of loans flagged as defaults that really defaulted: 40 true positives divided by 50 flagged loans. The value 0.67 would be recall, which divides by the 60 actual defaults rather than the flagged loans.
- A0.67
- B0.80Correct
- C0.20
- D0.95
Explanation
Precision = true positives / predicted positives = 40/50 = 0.80. The 0.67 figure is recall (40/60), which uses actual defaults as the base instead of flagged loans. Check: false positives = 10, so 40/(40+10) = 0.80.
Did you get it right without looking?
One question tells you little. A timed set on Machine Learning and Prediction shows your real accuracy, how long you take and where you lose marks.
More Machine Learning and Prediction questions
- A data scientist fits a model with very low training error but much higher error on a held-out validation set. Which is the most appropriate…
- A PCA on six standardized variables yields eigenvalues of 3.0, 1.5, 0.6, 0.45, 0.30 and 0.15. What is the minimum number of components neede…
- A risk analyst fits a single, very deep classification tree to predict loan default. It classifies the training data almost perfectly but pe…
- A bank tests a default-prediction classifier on 200 loans. The model flags 40 loans as defaults, of which 30 actually defaulted. In total 50…
- A lender lowers the probability threshold used to classify a borrower as a predicted defaulter, from 0.50 to 0.30, using the same fitted mod…
- A neural network with many parameters achieves very low error on the training set but substantially higher error on a validation set. Which …