Skip to content

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

A K-means model is fitted to bond data with features of duration (range 1 to 20 years) and bond price volatility expressed in basis points (range 50 to 5,000). No preprocessing is applied. What is the most likely problem?

The volatility feature will dominate. K-means relies on Euclidean distance, which is scale-sensitive, so the variable with the much larger numeric range (basis points) drives cluster assignments. Standardizing features before clustering prevents this distortion.

  1. AThe algorithm will fail to converge because the features have different units
  2. BThe volatility feature will dominate the Euclidean distance, so clusters mainly reflect volatilityCorrect
  3. CDuration will dominate because it is measured in years
  4. DK-means will automatically rescale features, so there is no problem

Explanation

K-means uses Euclidean distance, which is sensitive to scale. The feature with the much larger numeric range contributes most of the squared distance, so the basis-point volatility dominates. Standardizing the features fixes this.

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