What is forecasting?

Before we move on, let's make a distinction between two related terms. Time series analysis is the description, analysis, and/or search for insights and meaning of time series data that has already happened. Forecasting, as we've already read in the preface, is the prediction of future values of a discrete-time stochastic process.

We will, incidentally, be doing some time series analysis in the process of forecasting future values but, by virtue of being in the predictive analytics unit of this book, we will mainly be focused on prediction, rather than analysis proper.

One other note: When we speak of forecasting in this chapter, we are referring specifically to quantitative forecasting as opposed to judgemental or qualitative forecasts (which seek to make predictions in spite of a lack or dearth of historical data, or in light of an unforeseen "shock to the system".

In contrast, quantitative forecasting is used when we have sufficient numerical values for past observations...