We recommed that the vignettes are read and gone through the order as
shown here below. This is done such that the reader understand what the
package is doing and its capability. Also, that the data and model is
setup in the correct way.
There are four vignettes included in the package:
setup-data (code) covers how data must be setup
for an onlineforecast model. The vignette goes into details on how the
observations and the model inputs (predictors) are setup. The vignette
will also focus on the importance on understand the align the forecast
in correct time.
setup-and-use-model (code) focus on how to setup
the onlineforecast model and use it to generate forecasts.
model-selection (code) demonstrates how model
selection can be carried out.
forecast-evaluation (code) provides an
introduction to evaluation of generated forecasts and to analysis of the
results with tips on learning how to improve a model.
online-updating (code) demonstrates how to update
a model and forecasts when new observations becomes available.
- nice-tricks (code) provides useful some tips on
how to make the work flow easier with the package.
For more particular examples of forecast applications, see the examples. For an in-depth description, read onlineforecasting.