The Weighted Average Percent Error (WAPE) is a statistical measure used to evaluate the accuracy of forecasts, particularly in time series data, by calculating the total absolute error normalized by the total actual demand or sales.
Understanding Weighted Average Percent Error (WAPE)
WAPE helps to overcome limitations present in other error metrics, especially when dealing with different scales or magnitudes of data. It provides a single percentage value representing the overall forecast accuracy.
Formula for WAPE
The formula for calculating WAPE is:
WAPE = (Sum of Absolute Errors / Sum of Actual Values) * 100
Expressed formally:
WAPE = (Σ |Actual Value - Forecasted Value|) / (Σ |Actual Value|) * 100
Interpretation of WAPE
- A lower WAPE indicates a higher degree of forecast accuracy.
- WAPE is expressed as a percentage, making it easily understandable and comparable across different datasets.
Advantages of Using WAPE
- Scale-Independent: WAPE is useful when comparing forecasts across different scales or units.
- Easy to Interpret: Expressed as a percentage, it is easily understandable by non-technical stakeholders.
- Considers Magnitude: WAPE takes into account the magnitude of the actual values, providing a more balanced measure of error.
Example Calculation
Let's say you have the following data:
| Item | Actual Value | Forecasted Value | Absolute Error |
|---|---|---|---|
| A | 100 | 90 | 10 |
| B | 200 | 210 | 10 |
| C | 150 | 140 | 10 |
- Sum of Absolute Errors: 10 + 10 + 10 = 30
- Sum of Actual Values: 100 + 200 + 150 = 450
- WAPE: (30 / 450) * 100 = 6.67%
Therefore, the WAPE for this example is 6.67%.
Conclusion
WAPE is a valuable metric for assessing forecast accuracy, offering a balanced and easily interpretable measure of error that considers the scale and magnitude of the data. It is particularly useful in time series forecasting and demand planning.
