WMAPE

WMAPE (sometimes spelled wMAPE) stands for Weighted Mean Absolute Percentage Error.[1] It is a measure of prediction accuracy of a forecasting method. It's a variant of MAPE in which errors are weighted by sales volume.[1]

Its formula is:[2]

${\displaystyle {\mbox{WMAPE}}={\frac {\sum _{t=1}^{n}\left|A_{t}-F_{t}\right|}{\sum _{t=1}^{n}\left|A_{t}\right|}}}$

where ${\displaystyle A}$ is a vector of the actual data and ${\displaystyle F}$ is the forecast. The advantage of this metric over MAPE is that this overcomes the 'infinite error' issue.[2]