You may opt to treat seasonality in two basic ways in Statistical Forecasting, i.e.
- Use previously created seasonal factors for any group of items (Group factors). Group factors normally produce the most reliable results as they are not influenced so much by random fluctuations in actual demand for an individual item.
- Or, use average seasonal factors computed monthly variations of actual demand from the fitted trend model for the item concerned (Item Specific factors).
You may use any existing set of factors, including non-seasonal factors. Note that seasonal factors are used for two purposes:
- Removal of seasonality from actual demand before fitting the selected trend model to historical data
- Adjusting projected trend values in order to calculate monthly forecast values
See Seasonal Factor Estimation for more details.
Please sign in to leave a comment.