sieve_default {dwp}R Documentation

Test Criteria for Model Selection

Description

Test Criteria for Model Selection

Usage

sieve_default

Format

A list containing the parameters used for test criteria in model selection an ddArray objects. The sieve_default values are used as a default in modelFilter. If desired, users may create their own tests, using sieve_default as a template. The same list elements must all be present and have the same structure as the defaults, namely:

$aic

the cutoff for DeltaAIC scores; models with higher scores are removed from further consideration. Default is $aic = 10

$hin

a boolean to indicate whether or not to use high leverage points as a criterion for model selection. Default is $hin = TRUE

$rtail

a vector of probabilities that the fitted model must exceed at 80, 120, 150, and 200 meters. Default is rtail = c(p80 = 0.50, p120 = 0.90, p150 = 0.95, p200 = 0.99). Custom test parameters must be a vector probabilities with "p80", "p120", "p150", and "p200" in the names.

ltail

a vector of probabilities that a fitted model must not exceed at 20 and 50 meters. Default is ltail = c(p20 = 0.50, p50 = 0.90). Custom test parameters must be a vector of probabilities with "p20" and "p50" in names.


[Package dwp version 1.1 Index]