EMC2 (version 2.1.0)

compare_subject: Information criteria for each participant

Description

Returns the BPIC/DIC based model weights for each participant in a list of samples objects

Usage

compare_subject(
  sList,
  stage = "sample",
  filter = 0,
  use_best_fit = TRUE,
  print_summary = TRUE,
  digits = 3
)

Value

List of matrices for each subject of effective number of parameters, mean deviance, deviance of mean, DIC, BPIC and associated weights.

Arguments

sList

List of samples objects

stage

A string. Specifies which stage the samples are to be taken from "preburn", "burn", "adapt", or "sample"

filter

An integer or vector. If it's an integer, iterations up until the value set by filter will be excluded. If a vector is supplied, only the iterations in the vector will be considered.

use_best_fit

Boolean, defaults to TRUE, use minimal likelihood or mean likelihood (whichever is better) in the calculation, otherwise always uses the mean likelihood.

print_summary

Boolean (defaults to TRUE) print table of results

digits

Integer, significant digits in printed table

Examples

if (FALSE) {
# Define a list of two (or more different models)
# Here the full model is an emc object with the hypothesized effect
# The null model is an emc object without the hypothesized effect
design_full <- design(data = forstmann,model=DDM,
                           formula =list(v~0+S,a~E, t0~1, s~1, Z~1, sv~1, SZ~1),
                           constants=c(s=log(1)))
# Now without a ~ E
design_null <- design(data = forstmann,model=DDM,
                           formula =list(v~0+S,a~1, t0~1, s~1, Z~1, sv~1, SZ~1),
                           constants=c(s=log(1)))

full_model <- make_emc(forstmann, design_full)
full_model <- fit(full_model, cores_for_chains = 1)

null_model <- make_emc(forstmann, design_null, cores_for_chains = 1)
null_model <- fit(null_model)
sList <- list(full_model, null_model)
compare_subject(sList)
# prints a set of weights for each model for the different participants
# And returns the DIC and BPIC for each participant for each model.
}