Given x and y performs the MCMC optimization.
ltm_mcmc(x, y, burnin = 2000, iter = 8000, K = 3, prior_par = create_prior_parameters())
| x | data points |
|---|---|
| y | response variable |
| burnin | number of burnin iterations |
| iter | number of iterations after burnin |
| K | parameter K |
| prior_par | List of parameters for prior distrributions.
See |
matrix containing the posterior samples. Each line is one sample after the burnin period and each column is one of the parameters of the model. Columns are named to find the parameters with ease.
Nakajima, Jouchi, and Mike West. "Bayesian analysis of latent threshold dynamic models." Journal of Business & Economic Statistics 31.2 (2013): 151-164.
# Generates 10 series, each one with 500 observations and 2 regressors. d_sim <- ltm_sim( ns = 500, nk = 2, ni = 10, vmu = matrix(c(.5,.5), nrow = 2), mPhi = diag(2) * c(.99, .99), mSigs = c(.1,.1), dsig = .15, vd = matrix(c(.4,.4), nrow = 2), alpha = 0 ) # Fit model fit_model <- ltm_mcmc(d_sim$mx, d_sim$vy, burnin = 0, iter = 2)#> Iteration: 1 / 2 [ 50%] (Sampling) #> Iteration: 2 / 2 [ 100%] (Sampling)