Obtain partial effects of trained model

Create a partial response or effect plot of a trained model.

Usage

partial(
  mod,
  x.var = NULL,
  constant = NULL,
  variable_length = 100,
  values = NULL,
  newdata = NULL,
  plot = FALSE,
  type = "response",
  ...
)

# S4 method for class 'ANY'
partial(
  mod,
  x.var = NULL,
  constant = NULL,
  variable_length = 100,
  values = NULL,
  newdata = NULL,
  plot = FALSE,
  type = "response",
  ...
)

partial.DistributionModel(mod, ...)

Arguments

mod

A trained DistributionModel object with fit_best model within.

x.var

A character indicating the variable for which a partial effect is to be calculated.

constant

A numeric constant to be inserted for all other variables. Default calculates a mean per variable.

variable_length

numeric The interpolation depth (nr. of points) to be used (Default: 100).

values

numeric Directly specified values to compute partial effects for. If this parameter is set to anything other than NULL, the parameter "variable_length" is ignored (Default: NULL).

newdata

An optional data.frame with provided data for partial estimation (Default: NULL).

plot

A logical indication of whether the result is to be plotted?

type

A specified type, either 'response' or 'predictor'. Can be missing.

...

Other engine specific parameters.

Value

A data.frame with the created partial response.

Details

By default the mean is calculated across all parameters that are not x.var. Instead a constant can be set (for instance 0) to be applied to the output.

See also

partial

Examples

if (FALSE) { # \dontrun{
 # Do a partial calculation of a trained model
 partial(fit, x.var = "Forest.cover", plot = TRUE)
} # }