Chapter 17 Forecasting with ADAM
Finally, we come to the technicalities of how to produce forecasts using ADAM. We have already discussed in previous section how conditional expectations can be generated from some of the models (e.g. Sections 5.3, 6.3, 9.2.1, 10.2 and 10.3.1), but we have not discussed this in necessary detail. Furthermore, as discussed in Section 1.1 that forecasts should align with specific decisions, but we have not discussed how to do that.
In this chapter, we start the discussion with the principles behind calculating the conditional expectations from ADAM models (including ETS, ARIMA, regression and their combinations). We then move to the discussion of various methods for prediction intervals construction, starting from the basic parametric ones and ending with empirical and those that take uncertainty of parameters into account (see Section 16). Finally, we discuss prediction intervals for intermittent state space model (Section 13), one-sided intervals and cumulative forecasts over the forecast horizon, something that is useful in practice, especially when inventory decisions need to be made. We also discuss confidence interval for the conditional mean construction, the topic which is not as important as the others mentioned above, but which is still worth mentioning.