das saisonale box-jenkins modell

summary: the box-jenkins-method for estimating and forecasting times series models relies heavily on the interactive model building philosophy of identification, estimation and diagnostic checking. for practical purposes the identification and diagnostic checking phase are important. on seasonal box-jenkins-models (sarima-processes) there is only few literature. especially one has to know, which autocorrelation function (acf) leads to which models, and what are the main differences between the various types of seasonal models. to make the results practicable for the application of model building, a computer program for generation of acf and partial acf was developed and implemented in the iaz-system. the paper contains a general discussionof seasonal models and lists all important types of seasonal moving average models up to three parameters. for each model the theoretical acf and the graphical display of the acf is given. also there is a short discussion of the main features of each model, and arguments are given for the choice between different models.;