JointAI (version 0.6.1)

default_hyperpars: Get the default values for hyperparameters

Description

This function returns a list of default values for the hyperparameters.

Usage

default_hyperpars()

Details

norm: hyperparameters for normal and lognormal models

mu_reg_norm	mean in the priors for regression coefficients
tau_reg_norm	precision in the priors for regression coefficients
shape_tau_norm	shape parameter in Gamma prior for precision of an imputed variable

gamma: hyperparameters for Gamma models

mu_reg_gamma	mean in the priors for regression coefficients
tau_reg_gamma	precision in the priors for regression coefficients
shape_tau_gamma	shape parameter in Gamma prior for precision of an imputed variable

beta: hyperparameters for beta models

mu_reg_beta	mean in the priors for regression coefficients
tau_reg_beta	precision in the priors for regression coefficients
shape_tau_beta	shape parameter in Gamma prior for precision of imputed variable

logit: hyperparameters for logistic models

mu_reg_logit

mean in the priors for regression coefficients

probit: hyperparameters for probit models

mu_reg_logit

mean in the priors for regression coefficients

multinomial: hyperparameters for multinomial models

mu_reg_multinomial

mean in the priors for regression coefficients

ordinal: hyperparameters for ordinal models

mu_reg_ordinal	mean in the priors for regression coefficients
tau_reg_ordinal	precision in the priors for regression coefficients
mu_delta_ordinal	mean in the prior for the intercepts

Z: function creating hyperparameters for the random effects in mixed models, with output elements

RinvD	scale matrix in Wishart prior (*) for random effects covariance matrix
KinvD	degrees of freedom in Wishart prior for random effects covariance matrix
shape_diag_RinvD	shape parameter in Gamma prior for the diagonal elements of RinvD

(*) when there is only one random effect a Gamma distribution is used instead of the Wishart and RinvD and KinvD are NULL

surv: parameters for survival models (parametric and proportional hazard)

mu_reg_surv

mean in the priors for regression coefficients

coxph: parameters for Cox proportional hazards models

c confidence in prior guess for the hazard function
r failure rate per unit time

Examples

# NOT RUN {
default_hyperpars()

# To change the hyperparameters:
hyp <- default_hyperpars()
hyp$norm['rate_tau_norm'] <- 1e-3
mod <- lm_imp(y ~ C1 + C2 + B1, data = wideDF, hyperpars = hyp, mess = FALSE)


# }