Introduction
If we need to predict future values for, let's say oil price, we could model the price in terms of other variables (the number of rigs available, the number of companies investing in oil, and so on). These are usually referred to as cross-sectional models.
In the 1970s, many economists realized that these models were ineffective for predicting future values, and a new approach was proposed. The idea was then to use the series' past values to predict the future, in such a way that the model was chosen to mimic the temporal correlation structure as much as possible. For example, if a series had a strong serial correlation of order 1 (meaning that two consecutive values would usually be correlated), we would choose a model that ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access