Chapter 11. Predicting Changes with Time
The inventory and staffing of your favorite bookstore, the telecommunications route of every important phone call you make or website you visit, policy changes with regard to climate change, your retirement portfolio, exploitative pricing models that capitalize on consumer need and scarcity under the guise of efficient allocation of resources; all of these, and innumerable others, have one thing in common. They make heavy use of the statistical approach to predicting changes with time, or time series forecasting, based on data from past and present.
In this chapter, we are going to go through different methods of forecasting but only one overarching model (exponential smoothing/ETS), though there exist many others (ARIMA, RBF neural networks, and so on). There are a few reasons for this:
- First, the model we'll be talking about, as we'll see, is very widely applicable to a range of different time series with a range of different properties
- Second, concentrating...
