Statistical forecasting methods and features
Several statistical forecast methodologies can be implemented in SAP IBP and are currently available in the application. Each forecast methodology contains at least two of the listed features below:
- Baseline
- Irregular fluctuations
- Trend
- Cyclicity
- Seasonality
Statistical forecast methodologies:
- Regression (linear, logarithmic, exponential, power)
- Historical average/moving average (single, double)
- Driver-based forecast
- Exponential smoothing (simple, double, triple)
- Holt’s linear trend
- Decomposition (additive, multiplicative/linear/logarithmic, exponential, power)
- Winter’s additive/mulitplicative
Solution features
- Features of the automated statistical forecast application
- Forecast methodologies with associated parameters.
- Forecast scenarios for comparison.
- Forecast features adapt to forecast needs.
- Selecting forecast methods based on maximizing performance and minimizing variance.
- Visualization of all forecasting techniques with the possible parameters.
- Visualization of the optimal parameters vs. the actuals.
- Comparison of loaded actuals vs. previous forecast scenarios (margin %) to anticipate GAPs that can be explained.
