Model Class Functions

Model class functions #

A full list of functions to be added here.

MethodDescription
covarianceDThe matrix D, the covariance matrix of the random effects
ZThe matrix Z, the design matrix of the random effects
chol_D()Generates the Cholesky decomposition of D
log_likelihoodFor a given vector $u$ calculates the multivariate Gaussian log-likelihood with zero mean and covariance D
simulate_re()Simulates a vector $u$
sparse()Choose whether to use sparse matrix methods
parametersThe parameters $\theta$
formulaThe random effects formula used to create D
update_parameters()Updates $\theta$ and related matrices
nngp()Sets or returns options for the nearest neighbour Gaussian process approximation
hsgp()Sets or returns options for the Hilbert space Gaussian process approximation
meanXThe matrix X, the design matrix of fixed effects
parametersThe parameters $\beta$
offsetThe optional model offset
formulaThe fixed effects formula used to create X
linear_predictor()Generates the linear predictor
update_parameters()Updates $\beta$ and related matrices
familyAn R family object
var_parThe scale parameter
fitted()Full linear predictor including random effects (either simulated or provided)
predict()Predictions from the model at new data values
sim_data()Simulates data from the model
Sigma()Generates $\Sigma$ (or an approximation)
information_matrix()Generates the information matrix
sandwich()Generates the robust sandwich matrix
kenward_roger()Small sample bias-corrected variance matrix of $\hat{\beta}$
partial_sigma()Generates matrices $\partial \Sigma / \partial \theta$ and $\partial^2 \Sigma / \partial \theta_i \partial \theta_j$
use_attenuation()Option for improving the approximation of $\Sigma$
power()Estimates the power
MCML()MCMC Maximum likelihood model fitting
LA()Laplace approximation model fitting
mcmc_sample()Sample $u$ using MCMC
w_matrix()Returns $diag(W)$
dh_deta()Returns $\partial h^{-1}(\eta) / \partial \eta$
log_gradient()Returns the gradient of the log likelihood with respect to either the random effects or $\beta$
marginal()Calculates marginal effects with different types of conditioning or averaging
comments powered by Disqus