Skip to content

FRM Part I · FRM Exam Part I · Machine-Learning Methods

A risk analyst fits a regression of credit spread changes on 40 candidate predictors using only 120 observations. The ordinary least squares model fits the training data almost perfectly but forecasts poorly out of sample. Which approach most directly addresses this problem by adding a penalty on the size of the coefficients?

Ridge or LASSO regularization is correct because these methods add a penalty on coefficient size, shrinking estimates and reducing variance. With many predictors relative to observations, OLS overfits, and the penalty improves out-of-sample forecasting performance, whereas adding predictors would worsen overfitting.

  1. AAdding more predictors to the model
  2. BUsing ridge or LASSO regularizationCorrect
  3. CIncreasing the number of training epochs
  4. DRemoving the intercept from the model

Explanation

The model is overfitted because the number of predictors is large relative to observations. Ridge and LASSO add a penalty on coefficient size, shrinking coefficients and reducing variance. Adding predictors would worsen overfitting, and the other options do not penalize coefficients.

Did you get it right without looking?

One question tells you little. A timed set on Machine-Learning Methods shows your real accuracy, how long you take and where you lose marks.

Practise Machine-Learning MethodsFree diagnostic mock

More Machine-Learning Methods questions