A pmwgs object with a limited number of samples of the Forstmann dataset.
sampled_forstmannA pmwgs object minus the data. A pmwgs opbject is a list with a specific structure and elements, as outlined below.
A character vector containing the model parameter names
The number of parameters in the model
The number of unique subject ID's in the data
A vector containing the unique subject ID's
A list that holds the prior for theta_mu (the model
parameters). Contains the mean (theta_mu_mean), covariance matrix
(theta_mu_var) and inverse covariance matrix
(theta_mu_invar)
The log likielihood function used by pmwg for model estimation
A list with defined structure containing the samples, see the Samples Element section for more detail
The pmwgs object is missing one aspect, the pmwgs$data element. In order to fully replicate the full object (ie to run more sampling stages) you will need to add the data back in, via sampled_forstmann$data <- forstmann
The samples element of a PMwG object contains the different types of samples
estimated by PMwG. These include the three main types of samples
theta_mu, theta_sig and alpha as well as a number of
other items which are detailed here.
samples used for estimating the model parameters (group level), an array of size (n_pars x n_samples)
samples used for estimating the parameter covariance matrix, an array of size (n_pars x n_pars x n_samples)
samples used for estimating the subject random effects, an array of size (n_pars x n_subjects x n_samples)
A vector containing what PMwG stage each sample was drawn in
The winning particles log-likelihood for each subject and sample
Mixing weights used during the Gibbs step when creating a new sample for the covariance matrix
The inverse of the last samples covariance matrix
The index of the last sample drawn