Description:
The forecasting of time series in goods management systems causes various problems that we identify and indicate possible solutions. The implementation of auxiliary information like promotional activities or calendar effects in forecasts using ARMA models and exponential smoothing methods may be difficult, especially if these effects did not yet occur in the past. The effects usually cause transient shocks that have to be corrected when using ARMAX models and that yield high forecast variances. These high variances have to be corrected when estimating the parameters. The calendar effects do not for all items have a fixed seasonal figure but may have a variable seasonal behaviour, e.g. depending on the weather. We have to identify and perhaps to eliminate outliers that would yield not optimal forecasts. In addition to these problems dealing with univariate time series, we introduce some similarity-measures for time series which are prerequisite for the multivariate analysis of time series.