ABSTRACT
A time series is a set of consecutive measurements of the same variable made at equally spaced intervals in time. Well-known examples of naturally occurring time series over different time frames include stock markets, sales figures, rainfall, and birth/death rates. The primary objective of time series analysis (TSA) is to mathematically describe the entire series as an equation, which then allows prediction or forecasting of its future values. In contexts like engineering and manufacturing, TSA is also a key component of what is known as “process control”, i.e. interventions to achieve safe, consistent and/or profitable performance levels based on historical data (Alwan & Roberts, 1988). I will follow most writers in using the terms “predict” and “forecast” interchangeably in this book. Some argue that there are nuanced differences between these terms, but the argument relates more to statistical theory and is of little consequence here.
