metamodel: Linear regression metamodeling

In dampack: Decision-Analytic Modeling Package

Description

This function estimates a linear regression metamodel for a given decision-analytic model by using the results of a probabilistic sensitivity analysis (PSA)

Usage

metamodel(
  analysis = c("oneway", "twoway", "multiway"),
  psa,
  params = NULL,
  strategies = NULL,
  outcome = c("eff", "cost", "nhb", "nmb", "nhb_loss", "nmb_loss", "nhb_loss_voi",
    "nmb_loss_voi"),
  wtp = NULL,
  type = c("linear", "gam", "poly"),
  poly.order = 2,
  k = -1
)

Arguments

analysis	either "oneway" or "twoway"
psa	psa object
params	string vector with the name(s) of the parameter of interest. Defaults to all.
strategies	vector of strategies to consider. The default (NULL) is that all strategies are considered.
outcome	either effectiveness ("eff"), cost ("cost"), net health benefit ("nhb"), net monetary benefit ("nmb"), or the opportunity loss in terms of NHB or NMB ("nhb_loss" and "nmb_loss", respectively). "nmb_loss_voi" and "nhb_loss_voi" are only used in internal function calls of metamodel within other VOI functions.
wtp	if outcome is NHB or NMB (or the associated loss), must provide the willingness-to-pay threshold
type	type of metamodel
poly.order	order of polynomial for the linear regression metamodel. Default: 2
k	the dimension of the basis used to represent the smooth term. The default depends on the number of variables that the smooth is a function of. k should not be less than the dimension of the null space of the penalty for the term (see null.space.dimension), but will be reset if it is. See choose.k for further information.

Details

The most important option is analysis, which can be either "oneway" or twoway. If analysis == "oneway", a separate metamodel is created for each combination of the parameters in params and strategies in strategies (by default, this is all strategies and parameters).

If analysis == "twoway", params must be a vector of two parameters, and a metamodel is created with these two parameters for each strategy in strategies.

Value

A metamodel object, which contains a list of metamodels and other relevant information.

See Also

predict.metamodel, make_psa_obj, owsa, twsa

dampack documentation built on May 31, 2021, 1:06 a.m.