Skip to content

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

An analyst uses a regularized model with a tuning parameter that controls complexity. As the penalty is increased substantially from a very small value, which change in the bias and variance of predictions is most likely?

Bias increases and variance decreases. A stronger penalty makes the model simpler and shrinks coefficients, so it fits the true relationship less closely, raising bias, while it becomes less sensitive to the specific training sample, lowering variance.

  1. ABias decreases and variance decreases
  2. BBias increases and variance increases
  3. CBias decreases and variance increases
  4. DBias increases and variance decreasesCorrect

Explanation

A larger penalty shrinks coefficients and makes the model simpler and less flexible. This restricts its ability to fit the true pattern, so bias rises. At the same time the fitted model becomes less sensitive to the particular training sample, so variance falls.

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