Create the prior parameters.

Source: R/mcmc.R

Define the priors parameters to be used with ltm_mcmc().

create_prior_parameters(a_mu0 = 0, a_s0 = 0.1, n0 = 6, S0 = 0.06,
  v0 = 6, V0 = 0.06, m0 = 0, s0 = 1, a0 = 20, b0 = 1.5)

Arguments

a_mu0

mean of alpha normal distribution.
a_s0

standard deviation of alpha's normal distribution.
n0

sig2 inverse gamma shape parameter.
S0

sig2 inverse gamma location parameter.
v0

sig_eta inverse gamma shape parameter.
V0

sig_eta inverse gamma location parameter.
m0

mu normal's mean parameter.
s0

mu normals standard deviation.
a0

a0 beta's shape parameter.
b0

a0 beta's location parameter.

Value

List containing the hyperparameters used to fit the model. The default parameters are the same of the simulation example of the paper.

Details

Considering the following priors:

  • alpha ~ N(mu0, s0)

  • sig2 ~ IG(n0/2, S0/2)

  • sig_eta ~ IG(v0/2, V0/2)

  • mu ~ N(m0, s0^2)

  • (phi+1)/2 ~ Beta(a0, b0)

References

Nakajima, Jouchi, and Mike West. "Bayesian analysis of latent threshold dynamic models." Journal of Business & Economic Statistics 31.2 (2013): 151-164.