Figure 2
A graph that illustrates the long training scheme. It shows that the training period initially spans 10 years and it is increased by one year after each validation step. The graph also shows how the validation window that covers always 12 years is pushed forward by the increasing training window. The graph also depicts the one-year out-of-sample period, which is pushed forward by the training und validation periods. The graph shows the first, second, third, and 22th (the last) validation steps as an illustration.

Long training scheme. The figure depicts the one-year horizon variant of the long training scheme. The data range from October 1974 to December 2017. The training period initially spans 10 years and increases by one year after each validation step. Each of the 22 validation steps delivers a new set of parameter estimates. Each validation window covers 12 years and is rolled forward with a fixed width, followed by one year of out-of-sample testing.

Close
This Feature Is Available To Subscribers Only

Sign In or Create an Account

Close

This PDF is available to Subscribers Only

View Article Abstract & Purchase Options

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

Close