remove_nonpar_func: Reduction of Nonparametric Terms

In markcus1/gagam: Genetic Algorithm for Generalized Additive Models

Description

Additional reduction of nonparametric terms after the genetic algorithm has finished. Internal use only.

Usage

remove_nonpar_func(
  best_ind,
  all_dataframe,
  Kvar,
  Kint,
  k,
  bs,
  family,
  method,
  optimizer,
  always_par
)

Arguments

best_ind	Vector contain the model.
all_dataframe	Data frame with all variables.
Kvar	Maximum number of variables allowed in the final model
Kint	Maximum number of interactions allowed in the final model
k	Basis dimension for nonparametric terms estimated using cubic splines.
bs	Spline basis.
family	Specifies the family for the gam (see ?family and ?family.mgcv). Default is gaussian().
method	Specifies the metric for smoothing parameter selection (see ?gam). Default is "REML".
optimizer	Specifies the numerical optimization algorithm for the gam (see ?gam). Default is c("outer","newton").
always_par	Specifies which variables are always estimated parametrically

Value

Reduced model.

markcus1/gagam documentation built on April 16, 2020, 11:50 a.m.