Skip to content

FRM Part I · FRM Exam Part I · Machine Learning and Prediction

A bank combines 100 decision trees by majority vote to classify counterparties as default or non-default. Compared with a single deep tree, which outcome is most expected from the ensemble?

The ensemble typically has lower variance and better out-of-sample accuracy than a single deep tree, but it is less interpretable because it is no longer one simple set of rules. Averaging does not remove bias entirely.

  1. ALower variance and typically better out-of-sample accuracy, but reduced interpretabilityCorrect
  2. BHigher variance and better interpretability
  3. CZero bias because averaging removes all error
  4. DIdentical predictions, since all trees use the same data

Explanation

Averaging or voting across many different trees reduces variance and usually improves generalization. The cost is that the ensemble cannot be read as a single simple rule set. Averaging does not eliminate bias or irreducible error.

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.

Practise Machine Learning and PredictionFree diagnostic mock

More Machine Learning and Prediction questions