summary.dynamic_fit: Summarise a fitted dynamic model

In DynCount: Bayesian Dynamic Models for Count Time Series

Summarise a fitted dynamic model

Description

Summarise a fitted dynamic model

Usage

## S3 method for class 'dynamic_fit'
summary(object, probs = c(0.025, 0.5, 0.975), ...)

Arguments

object	A "dynamic_fit" object.
probs	Quantile probabilities. Default c(0.025, 0.5, 0.975).
...	Unused.

Value

An object of class "summary.dynamic_fit" containing posterior summaries of the global parameters (params, one row per parameter) and of the fitted values (fitted). The rows of params and the draws they summarise are

innov_sd

sqrt(draws$innov_var): the estimated increment SD for "gaussian" innovations, the marginal SD for "t" and "mixture", and the root of the series-average variance for "sv".

t_df

draws$nu, the Student-t degrees of freedom.

sv_mu, sv_phi, sv_sigma

draws$sv_mu, draws$sv_phi and draws$sv_sigma, the parameters of the AR(1) log-variance process.

ar1_rho

draws$rho, the AR(1) coefficient.

drift_mu / intercept_mu

draws$mu, the random-walk drift or the AR(1) intercept.

gate_open_prob

draws$pi_open, the probability that the zero-inflation gate is open (one minus the structural-zero probability).

Rows are included only where relevant. For "mixture" innovations only innov_sd is reported. The mixture components are exchangeable and not identified individually, so per-component summaries are not given. The draws of the weights and component variances are in draws$mix_weight and draws$mix_var. For the multinomial family every latent parameter is reported once per non-baseline category, with the category label in square brackets (e.g. innov_sd[B]), and the fitted summary is a long data frame with a category column covering all K categories.

Examples

sim <- simulate_dynamic_poisson(50, 0.2, 2, seed = 1)
fit <- fit_dynamic_model(sim$y, nsave = 200, nburn = 100, seed = 1)
summary(fit)

DynCount documentation built on Sept. 28, 2026, 5:10 p.m.