Which metric expresses error as a percentage of actual values?

Get ready for the GARP Risk and AI Exam with flashcards and multiple choice questions. Each question comes with hints and explanations. Prepare for success!

Mean Absolute Percentage Error expresses error as a percentage of actual values. It evaluates each observation by taking the absolute difference between the actual value and the forecast, dividing by the actual value to get a relative error, multiplying by 100 to convert to a percentage, and then averaging these percentages across all observations. This makes the error scale-free, so you can compare performance across series with different magnitudes or units. Other common metrics like MAE report errors in the same units as the data, while MSE and RMSE involve squared errors and return results in squared or original units, not percentages. That’s why this percentage-based measure is the one that expresses error relative to actual values. Note that it can be undefined or unstable if actual values are zero or near zero, which is a caveat to keep in mind.