Which evaluation metric is most sensitive to large prediction errors in numeric prediction?

Question

BlackTom AI · Accepted Answer

When evaluating numeric predictions, different metrics respond differently to the size of errors.
Option 1: MAE (Mean Absolute Error) averages the absolute residuals. Large errors increase this metric, but their impact grows linearly, so it is not as sensitive to very large errors as quadratic measures.
Option 2: MAPE (Mean Absolute Percentage Error) scales errors by the true values, which can underweight or overweight errors depending on the scale of the target and can be undefined or inflated for small true values. Its sensitivity to outliers is not as direct or consistent as a squared-error metric.
Option 3: RMSE (Root Mean Squared Error) squares each residual, which disproportionately amplifies larger errors, then takes the square root. This makes RMSE highly sensitive to large prediction mistakes because a single large error can dominate the metric.
Option 4: ME (Mean Error) captures the average signed error, indicating bias but not the magnitude distribution. It can mask large individual errors by canceling positive and negative residuals, so it does not reflect the presence of large mistakes.
In summary, the squaring step in RMSE causes large errors to have a disproportionately large influence, relative to MAE, MAPE, and ME, making RMSE the most sensitive to substantial prediction mistakes.

Which evaluation metric is most sensitive to large prediction errors in numeric prediction?单项选择题

类似问题

Question at position 9 True/False Question: MSE penalizes large errors more than MAE. TrueFalse

Question at position 5 True/False Question: RMSE depends on the size of the test data. TrueFalse

The RMSE can be calculated using the following equation: RMSE=√∑ni=1(yi−ˆyi)2n While the bias can be calculated: BIAS=∑ni=1(yi−ˆyi)n Based on these equations, which of the following statements is TRUE?

A project has a required return of 12.6 percent, an initial cash outflow of $42,100, and cash inflows of $16,500 in Year 1, $11,700 in Year 2, and $10,400 in Year 4. What is the net present value?

更多留学生实用工具

智能学习助手

风格化写作助手

论文查重助手

文献引用助手

课堂转译助手

课堂笔记助手

Quiz搜索助手

学校历年真题

智能刷题助手

智能匹配练习

希望你的学习变得更简单