crossValidFunction: CROSS-VALIDATION FUNCTION

In CeresBarros/ToolsCB: Ceres's Useful R Functions

CROSS-VALIDATION FUNCTION

Description

CROSS-VALIDATION FUNCTION

Usage

crossValidFunction(
  fullDT,
  statsModel,
  origData,
  k = 4,
  idCol,
  parallel = FALSE,
  cacheObj1 = NULL,
  cacheObj2 = NULL,
  cacheArgs = NULL,
  level = NULL,
  ...
)

Arguments

fullDT	data.table with full dataset
statsModel	the statistical model to validate. Only works with gamlss models
origData	the data used to fit the statsModule, needs to be passed to gamlss::predictAll() (it may not be able to access it) but also to make sure newdata in gamlss::predictAll() has the same variables (even if they're not used in the model)
k	integer with number of chunks that the data should be partitioned in
idCol	column with pixel/observation IDs (optional)
parallel	logical. Uses future.apply::future_lapply() to parallelise model fitting across the k-folds, using plan(multiprocess). Defaults to FALSE
cacheObj1	object used by reproducible::Cache() for digesting, to avoid digesting the (potentially) large data arguments
cacheObj2	object used by reproducible::Cache() for digesting, to avoid digesting the (potentially) large data arguments
cacheArgs	a named list of arguments passed to inner Cache calls
level	passed to gamlss:::predict
...	further arguments passed to future::plan()

CeresBarros/ToolsCB documentation built on Aug. 23, 2024, 4:22 p.m.