Given the current parameters and a covariance matrix, returns a vector for a proposed jump from a multivariate normal distribution. Takes into account parameter covariance and ensures containment condition with beta, if cov_mat0 (the identity matrix) is specified.
mvr_proposal( values, fixed, cov_mat, cov_mat0 = NULL, use_log = FALSE, beta = 0.05 )
| values | the vector of current parameter values |
|---|---|
| fixed | set of flags corresponding to the parameter vector indicating which parameters are fixed Takes into account parameter covariance and ensures containment condition with beta, if cov_mat0 (the identity matrix) is specified. |
| cov_mat | the 2D covariance matrix for all of the parameters |
| cov_mat0 | optional, usually the identity matrix for theta |
| use_log | flag. If TRUE, propose on log scale |
| beta | Beta as in Rosenthal and Roberts 2009 |
a parameter vector of a proposed move. Note that these may fall outside the allowable ranges.
Other proposals:
inf_hist_swap_phi(),
inf_hist_swap(),
infection_history_symmetric(),
univ_proposal()