Skip to contents

Plot VAR predictions

Source: R/ggvar_predict.R
ggvar_predict.Rd

Gets predictions via stats::predict and plots them. Can also plot the known values of the past.

Usage

ggvar_predict(
  x,
  series = NULL,
  index_ahead,
  index_behind = NULL,
  ci = 0.95,
  args_aes = list(),
  args_line = list(),
  args_ribbon = list(fill = NA, linetype = 2, color = "blue"),
  args_labs = list(),
  args_facet = list(),
  ...
)

Arguments

x

A "varest" object to get predictions from.

series

A character vector with series (variables) to consider. Defaults to all (NULL).

index_ahead

A vector of labels to the x-axis, normally dates. Applied to the predicted portion of the graph. Its length will define the prediction horizon.

index_behind

A vector of labels to the x-axis, normally dates. Applied to the original portion of the graph. Its length will define the 'past' horizon. Leave as NULL to only plot predicted values.

ci

The confidence level for the confidence interval. Set to FALSE to omit. Used in predict.

args_aes

Defines aesthetics to differentiate the data. A named list of aesthetics names (*) – arguments passed to ggplot2::scale_*_manual. See more in the 'Customization' section.

args_line, args_ribbon

Additional arguments passed to geom_line and geom_ribbon (respectively). See more in the 'Customization' section.

args_labs

Additional arguments passed to labs. If an empty list, will be changed to default values.

args_facet

Additional arguments passed to the faceting engine used.

...

Arguments passed to methods, see the 'Methods' section.

Value

A ggplot.

Details

Customization

The graph can be customized both with the 'static' arguments passed to each layer – using the args_* arguments –, and, if applicable, the 'dynamic' aesthetics – using the args_aes argument.

The args_aes is a list with '* = arguments to scale_*_manual \ elements, where '*' represents the name of an aesthetic to apply to the \ data. View vignette('ggplot2-specs', 'ggplot2') to see the available \ aesthetics.

After built, the result can be further customized as any ggplot, adding or overwriting layers with the ggplot's +. It is useful to understand the data and the mappings coded by the package, using the function get_gg_info.

See vignette('customizing-graphs') for more details.

Methods

The data from x is extracted with the generic function mtsdesc:::predict_setup. Each class conditions an external function to pass the ... arguments to. Below there is a list with all the currently implemented classes:

See also

Other historic values plots: ggvar_fit(), ggvar_history()

Examples

x <- vars::VAR(freeny[-2])
ggvar_predict(x, NULL, 1:10, 0:-10, args_facet = list(scales = "free_y"))
#> Warning: Some aesthetic must have its 'values' defined in `args_aes`. Setting
#> `args_aes$color$values` to "'ggplot'".