Forecast Methods (whina2100m000)

Use this session to list and maintain the forecast methods.

For each user defined forecast method, you can choose one of the following methods:

  • Moving Average
  • Exponential Smoothing
  • Previous Year's Calculation
  • Last Period's Demand

Forecast method codes can be linked to items, in the Item - Warehousing (whwmd4600m000) session or to items and warehouse combinations in the Item Data by Warehouse (whwmd2110s000) session.

Field Information

Forecast Method Code

The code of the forecast method.

Forecast Method

forecast method

Number of Periods for Moving Average

Number of periods for calculating the moving average.

Minimum Number of Periods of History Demand

The minimum number of periods for calculating the history demand.

Number of Periods Backward in Previous Year

The number of periods backward in previous year used to calculate the forecast.

Note: You can only specify the Number of Periods Backward in Previous Year if the forecast method is Previous Year's Calculation.
Number of Periods Forward in Previous Year

The number of periods forward in the previous year used to calculate the demand forecast.

Note: You can only specify the Number of Periods Forward in Previous Year if the forecast method is Previous Year's Calculation.
Smoothing Factor for Demand Forecast

The smoothing factor for the demand forecast. You must specify the smoothing factor if the forecast method is Exponential Smoothing.

The smoothing factor for the demand forecast indicates to what extent the difference between actual issue and the forecast in a certain period is included in the calculation of the demand forecast.

The formula that is used by the forecast method Exponential Smoothing to calculate the demand forecast is as follows:

nf = pf + s (ai -pf)
  • nf - new forecast
  • pf - previous forecast
  • ai - actual issue of previous period
  • s - smoothing factor

A low smoothing factor results in a high degree of smoothing. The smoothing factor is at least 0.01 and not greater than 1. Usually, the smoothing factor is between 0.1 and 0.3.

Smoothing Factor for Forecast Error

The smoothing factor for the forecast error.

The smoothing factor for the forecast error is one of the factors in the formula for the calculation of the absolute and relative forecast error. The absolute and relative forecast errors are only calculated in the demand forecast calculation, if the forecast method is Exponential Smoothing.

The absolute forecast error is used in the calculation of the safety stock in the Calculate Demand Forecast (whina2202m000) session.

The absolute and relative forecast errors are used in the calculation of the tracking signal.

Note: If the forecast method is Exponential Smoothing, the Smoothing Factor for Forecast Error must be greater than or equal to 0.01.
Smoothing Factor

If this check box is selected, a signal is displayed when the smoothing factor is used. This field is only relevant in case of a demand forecast calculated using the forecast method Exponential Smoothing. Sudden changes in the demand will make forecasts lag behind the actual demand. A tracking signal will signal such deviations.

The tracking signal factor is calculated by dividing the relative forecast error by the absolute forecast error.

If the tracking signal factor is larger than the critical tracking signal factor entered, this is caused by a sudden change in demand. The Smoothing Factor for Forecast Error field will then be too low for the forecast to react to incidental disturbances. In that case the calculation must be based on the tracking signal factor rather than the Smoothing Factor for Forecast Error field.

As soon as the tracking signal factor is less than the critical tracking signal factor, the value in the Smoothing Factor for Demand Forecast field will be used again.

Critical Tracking Signal Factor

The minimum and critical factor for initiating the tracking signal.

This field is only relevant in case of a demand forecast calculated using the forecast method Exponential Smoothing. Sudden changes in the demand will make forecasts lag behind the actual demand. A tracking signal will signal such deviations.

The tracking signal factor is calculated by dividing the relative forecast error by the absolute forecast error.

If the tracking signal factor is larger than the critical tracking signal factor entered in the current field, this is caused by a sudden change in demand. The Smoothing Factor for Forecast Error field will then be too low for the forecast to react to incidental disturbances. In that case the calculation must be based on the tracking signal factor rather than the Smoothing Factor for Forecast Error field.

As soon as the tracking signal factor is less than the critical tracking signal factor, the value in the Smoothing Factor for Demand Forecast field is used again.

Related topics