Multiple regression configuration
Multiple regression requires some additional configuration steps. When this forecast method is selected, the number of weeks of history to use must be indicated in the location setup. Then, the relationship between drivers can be set up at this path:
This table shows the fields to configure for multiple regression mapping:
Field | Description | Values |
---|---|---|
Location | The dependent driver to forecast | |
Mapped Location | The independent drivers | |
Forecast Source | The data source. Records using POS History are used to derive the coefficient that determines the relationship between the independent and dependent drivers. | POS HISTORY or FORECAST IMPORT |
Coefficient Order | A numeric value that is used to line up each coefficient with the imported forecast data. | Typically 1, 2, 3, etc. Each dependent driver will have 2 records for each Coefficient Order number. |
Day Offset | Used to indicate whether the source is the same day (0) or the previous day (1). This Day Offset is provided in situations where transactions on the previous day affect the next day’s forecast. For example, last night’s occupancy rate may be used to predict today’s Breakfast sales. | 0 for same day 1 for previous day |
Each dependent driver must have 2 mapping records, but may have more, depending on how many independent drivers are desired in the multiple regression forecast.