Seasonal Linear Regression 

Use

You can use seasonal linear regression (forecast strategy 35) alternatively to forecast with season models (forecast strategy 30) or forecast according to Winters (forecast strategy 31). Use seasonal linear regression especially if the historical time series contains many zeros or very small values. Forecast strategy 30 could calculate basic values that are too high in these cases.

Features

Before the system applies seasonal linear regression, it carries out a seasonal test. The system uses this test to check if the historical data contains any seasonal patterns. For this, the system determines the autocorrelation coefficient for all periods. If the value determined is at least 0.3, the system applies seasonal linear regression. If the value is less than 0.3, the system does not recognize a seasonal pattern and applies linear regression.

You can use method PARAMETER_SET of Business Add-In (BAdI) /SAPAPO/SCM_FCSTPARA to change the limit of 0.3

Activities

The system calculates the seasonal linear regression line as follows:

...

       1.      The seasonal indexes are calculated:

Determination of the starting seasonal index for each historical period t

                            a.      The number n_k of seasons available within the whole historical time series is calculated:

n_{k =} n_{total /} n_season

where n_season is the number of periods per season and n_total is the total number of historical values.

                            b.      The average value A_k of each season k is calculated:

A_k = Σ V(t) / n_season

where V(t) is the historical value of period t and n_season is the number of periods per season.

                            c.      The starting seasonal index s_start (t) is calculated for each period t within each season.

S_start (t) = V(t) / A_k

If a non-completed season exists (that is, if n_k is not an integer number) the starting seasonal index s_start (t) of the oldest historical data is calculated with the average A_k of the n_kth season.

Determination of the average seasonal index

                            d.      If k complete seasons are available, the starting seasonal indexes are averaged:

s_average(s) = (s_start (s) + s_start(n_season + s) + ....+s_start((k-1) n_season+s))/k,

s = 1, ... , n_season

Smoothing of the average seasonal indexes

                            e.      The result from step d is smoothed by a specified smoothing factor. The recommended value for the smoothing factor, which is set by default, is 1. You can, however, change this value by using the PARAMETER_SET method in the BAdI /SAPAPO/SCM_FCSTPARA.

       2.      The actual data is corrected on the basis of the seasonal indexes calculated in step 1.

       3.      Linear regression is performed on the non-seasonal actual values.

       4.      The seasonal indexes are applied to the results of the linear regression calculation, which produces the forecast results.