Broadly, a time series is a set of data points, for one or more variables, that is put in the order of the time they occurred. When we speak of time series in this chapter, we are referring specifically to regularly spaced, fixed interval time series. As such, the measurements can be taken yearly, every second, and everything in between and beyond, as long as there is an equal interval of time between each successive observation and measurements exist at the end of every interval.
For the purposes of statistical time series forecasting, we treat observations as realizations of random variables, much like in virtually everything we've done so far. Pointedly, observations of a time series are realizations of astochastic process. Because our observations are recorded in intervals, as opposed to continuously, you can refer to time series data as adiscrete-timestochastic process,to impress a date (update: don't do this, it will backfire)!