Process models of time series
Since very often we observe random data, in order to predict it, we need to find the most suitable model that would describe the behavior of this data. A time series model that uses random variables is called a process. Thus, if a time series is a sequence which is strongly known to us (for example, as a result of observing), then the time series process is a random time series, and its values will be different every time depending on the values that the random magnitudes take.
There are several models of time series processes. In order to understand which model is most suitable for sampling data, it is necessary to explore each of them. Next, when we know that the time series refers to a specific type, we can compute the estimates for the model parameters and make a forecast on this basis. For this reason, let's review these models one after another and see how they are implemented in Mathematica.
The moving average model
The moving average model is specified...