Chapter 6 Pure multiplicative ADAM ETS
There is a reason why we discuss pure multiplicative ADAM ETS models separately: they are suitable for the positive data, especially when the level is low, yet they do not rely on prior data transformations (such as taking logarithms or applying a power transform), which makes them useful in a variety of contexts. However, the models discussed in this chapter are not easy to work with – they typically do not have closed forms for the conditional \(h\) steps ahead mean and variance and do not have well-defined parameter space. Furthermore, they make more sense in conjunction with positive-valued distributions, although they also work with the Normal one. All these aspects are discussed in this chapter.
The software behind this book. The methods described here are implemented in the smooth and greybox packages for R and Python, which are free and open source.
Want to learn this with us? We teach these methods to practitioners on Demand Forecasting Principles, a four-week online course, and run other courses in forecasting, statistics and analytics.