run_cvfun: Create 'cvfits' from 'cvfun'

In projpred: Projection Predictive Feature Selection

Create cvfits from cvfun

Description

A helper function that can be used to create input for cv_varsel.refmodel()'s argument cvfits by running first cv_folds() and then the reference model object's cvfun (see init_refmodel()). This is helpful if K-fold CV is run multiple times based on the same K reference model refits.

Usage

run_cvfun(object, ...)

## Default S3 method:
run_cvfun(object, ...)

## S3 method for class 'refmodel'
run_cvfun(
  object,
  K = if (!inherits(object, "datafit")) 5 else 10,
  seed = NA,
  ...
)

Arguments

object	An object of class refmodel (returned by get_refmodel() or init_refmodel()) or an object that can be passed to argument object of get_refmodel().
...	For run_cvfun.default(): Arguments passed to get_refmodel(). For run_cvfun.refmodel(): Currently ignored.
K	Number of folds. Must be at least 2 and not exceed the number of observations.
seed	Pseudorandom number generation (PRNG) seed by which the same results can be obtained again if needed. Passed to argument seed of set.seed(), but can also be NA to not call set.seed() at all. If not NA, then the PRNG state is reset (to the state before calling run_cvfun()) upon exiting run_cvfun().

Value

An object that can be used as input for cv_varsel.refmodel()'s argument cvfits.

Examples

# Data:
dat_gauss <- data.frame(y = df_gaussian$y, df_gaussian$x)

# The "stanreg" fit which will be used as the reference model (with small
# values for `chains` and `iter`, but only for technical reasons in this
# example; this is not recommended in general):
fit <- rstanarm::stan_glm(
  y ~ X1 + X2 + X3 + X4 + X5, family = gaussian(), data = dat_gauss,
  QR = TRUE, chains = 2, iter = 500, refresh = 0, seed = 9876
)

# Define the reference model object explicitly:
ref <- get_refmodel(fit)

# Run the reference model object's `cvfun` (with a small value for `K`, but
# only for the sake of speed in this example; this is not recommended in
# general):
cvfits <- run_cvfun(ref, K = 2, seed = 184)

# Run cv_varsel() (with L1 search and small values for `nterms_max` and
# `nclusters_pred`, but only for the sake of speed in this example; this is
# not recommended in general) and use `cvfits` there:
cvvs_L1 <- cv_varsel(fit, method = "L1", cv_method = "kfold",
                     cvfits = cvfits, nterms_max = 3, nclusters_pred = 10,
                     seed = 5555)
# Now see, for example, `?print.vsel`, `?plot.vsel`, `?suggest_size.vsel`,
# and `?ranking` for possible post-processing functions.

# The purpose of run_cvfun() is to create an object that can be used in
# multiple cv_varsel() calls, e.g., to check the sensitivity to the search
# method (L1 or forward):
cvvs_fw <- cv_varsel(fit, method = "forward", cv_method = "kfold",
                     cvfits = cvfits, nterms_max = 3, nclusters = 5,
                     nclusters_pred = 10, seed = 5555)

projpred documentation built on Oct. 1, 2023, 1:07 a.m.