H2ALL: Empirical determinacy estimates based on the squared...

In hunansona/ed4bhm: Empirical Determinacy for Bayesian Hierarchical Models

Empirical determinacy estimates based on the squared Hellinger distance (H2)

Description

This function calculates a matrix with columns representing empirical determiancy measure estimates for all squared Hellinger (H2) distance types of five measures, which are TED, EDL, EDS, pEDL and pEDS defined in Hunanyat et al., 2021.

Usage

H2ALL(descriptive.matrix.ls)

Arguments

descriptive.matrix.ls

is a list composed of the descriptive matrix and delta δ

Details

The descriptive matrix has six columns containing means and standard deviations for minus, base and plus model fits and δ is the numerical step with w = 1-δ and w = 1 + δ being the factor to weight the likelihood. descriptive.matrix.ls must be an output of extract_descriptives_mbp_inla or extract_descriptives_mbp_jags.

The columns of the returned matrix are called c("TED", "EDL", "EDS", "pEDL", "pEDS") for the corresponding measures defined in Hunanyan et al., 2021. The rows of this matrix correspond to model parameters.

Value

numeric matrix with ncol = 5

References

Hunanyan, S., Roos, M., Plummer, M., Rue, H. (2021). Quantification of empirical determinacy: the impact of likelihood weighting on posterior location and spread in Bayesian meta-analysis estimated with JAGS and INLA. Bayesian Analaysis (in preparation)

See Also

extract_descriptives_mbp_inla, extract_descriptives_mbp_jags.

hunansona/ed4bhm documentation built on June 15, 2022, 6:42 p.m.