stan_fit_convert: Convert Stan Fit to Array of Samples

In tsnet: Fitting, Comparing, and Visualizing Networks Based on Time Series Data

Convert Stan Fit to Array of Samples

Description

This function converts a Stan fit object into an array of samples for the temporal coefficients and the innovation covariance or partial correlation matrices. It supports rstan as a backend. It can be used to convert models fit using stan_gvar into 3D arrays, which is the standard data structure used in tsnet. The function allows to select which parameters should be returned.

Usage

stan_fit_convert(stan_fit, return_params = c("beta", "sigma", "pcor"))

Arguments

stan_fit	A Stan fit object obtained from rstan or a tsnet_fit object from stan_gvar.
return_params	A character vector specifying which parameters to return. Options are "beta" (temporal network), "sigma" (innovation covariance), and "pcor" (partial correlations). Default is c("beta","sigma", "pcor").

Value

A list containing 3D arrays for the selected parameters. Each array represents the posterior samples for a parameter, and each slice of the array represents a single iteration.

Examples

data(ts_data)
example_data <- ts_data[1:100,1:3]
fit <- stan_gvar(data = example_data,
                 n_chains = 2,
                 n_cores = 1)
samples <- stan_fit_convert(fit, return_params = c("beta", "pcor"))

tsnet documentation built on June 20, 2025, 9:08 a.m.