lio.clogit: Leave-individuals-out for SSF

In ABoVE-AotM/above: Functions and methods for the Animals on the Move project of the Arctic Boreal Vulnerability Experiment (ABoVE - NASA).

Description

A function for generating leave-individuals-out cross-validation output for a step-selection model.

Usage

lio.clogit(
  dat,
  form,
  thin = NULL,
  cluster = "ID_year",
  ID = "AnimalID",
  inPar = TRUE,
  ncores = NULL
)

Arguments

dat	a movetrack object with pseudo-absences points (.e., available; Used = 0) and used points (Used = 1) stratified by step and annotated with representative covariates.
form	a formula object with the clogit formula to be evaluated using LIO cross-validation.
thin	an integer representing the amount of thinning required before performing the cross-validation (not recommended and potentially biases results).
cluster	a character string for the name of the clustering variable (e.g., 'ID_year').
ID	a character string for the name of the animal IDs. Used to partition individuals for the LIO procedure.
inPar	Boolean for whether or not procedure should be performed in parallel (TRUE by default).
ncores	if inPar = TRUE, ncores will define the number of cpu cores to use.

Value

Returns a list of lists indexed by individual. Nested within each individual's element is a second list containing the fitted and observed values (obj[[i]]$values) and cross-validation metrics (obj[[i]]$cv).

Author(s)

Peter Mahoney

See Also

processMovedata

ABoVE-AotM/above documentation built on May 28, 2020, 6:08 a.m.