compute_information_criteria: Compute the AIC, DIC and WAIC

In nathoze/Rsero: Estimate the annual force of infection using serological data

Compute the AIC, DIC and WAIC

Description

This function returns the Aikake Information Criterion (AIC), the Deviance Information Criterion (DIC) and the Watanabe Aikake Information Criterion (WAIC) from a fit of the class FOIfit.

Usage

compute_information_criteria(FOIfit, ...)

Arguments

FOIfit

A FOIfit object

Value

A list with the class information_criteria, which contains the following items:

• AIC: The Aikake Information Criterion.

• DIC: The Deviance Information Criterion.

• WAIC: The Wakanabe-Aikake Information Criterion.

• MLE: The Maximum-likelihood, estimated as the maximal value of the likelihood in the drawn samples, used in the AIC.

• k: Number of parameters, used in the AIC.

• Dbar: Mean deviance, used in the DIC.

• pD: Effective number of parameters of the model, used in the DIC.

• pwaic: Estimated effective number of parameters, used in the WAIC.

• lpd: log pointwise predictive density, used in the WAIC.

Author(s)

Nathanael Hoze [e-mail]

References

See Gelman et al., Understanding predictive information criteria for Bayesian models. Stat Comput (2014) 24:997-1016, DOI 10.1007/s11222-013-9416-2

Examples

data('one_peak_simulation')
model <- FOImodel(type='outbreak', K=1)
F1  = fit(model = model, data = data )
compute_information_criteria(FOIfit  = F1)

nathoze/Rsero documentation built on Feb. 3, 2024, 9:58 p.m.