Time series data is structurally different from the data discussed up until now. A glimpse of time series data was seen in Overseas Visitor in section 1 of Chapter 1, Introduction to Ensemble Techniques, and the bootstrapping of the time series models was briefly touched on in Chapter 2, Bootstrapping. The complexity that arises in the analysis of time series data is that the observations are not independent and, consequently, we need to specify the dependence. Box et al. (2015) is the benchmark book for the statistical analysis of time series, and its first edition was published in 1970. The class of models invented and popularized in Box and Jenkins is the popular autoregressive integrated moving average, famously abbreviated as ARIMA. This is also often known as the Box-Jenkins model.

Table 1 summarizes twenty-one time series datasets. The Length column gives the number of observations/data points of the series, while the Frequency column gives the periodicity of...