Skip to content

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

A credit-scoring team has 200 candidate predictors, many highly correlated in groups, and wants both variable selection and stable handling of correlated groups. Which method is most appropriate, and why?

Elastic net is most appropriate. Its combined L1 and L2 penalties let it set some coefficients to zero for selection while the L2 part stabilizes estimates and tends to keep correlated predictors together. LASSO alone often picks one arbitrarily, and ridge cannot select variables.

  1. AElastic net, because combining L1 and L2 penalties selects variables while tending to keep correlated predictors togetherCorrect
  2. BPure LASSO, because it always retains all members of a correlated group
  3. CPure ridge, because it sets irrelevant coefficients exactly to zero
  4. DOrdinary least squares, because correlated predictors do not affect it

Explanation

Elastic net mixes the L1 penalty (selection) and L2 penalty (stability under correlation, grouping effect). Pure LASSO tends to pick one variable from a correlated group arbitrarily. Ridge does not perform selection, and OLS becomes unstable with highly correlated predictors.

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