Because of the decomposition of the beta covariance matrix into scale and correlation, it can be hard to interpret the posteriors. This function will return an array which is a T-by-T-by-Samples array (where T is the number of coefficients). It is then easy to use apply to get posterior means for each cell in the matrix (e.g., apply(Sigma_Array, c(1, 2), mean)).

extract_cor_cov_samps(splt_fit, par_subscript = "ep")

Arguments

splt_fit

a fit of class stanfit

par_subscript

the subscript for the L_Omega_subscript and tau_subscript parameters

Value

a list with both Omega (correlation matrix) and Sigma (covariance matrix) sample arrays.