priors.spec: Specify the priors. Without inputs, defaults will be used.

In DGM: Dynamic Graphical Models

Description

Specify the priors. Without inputs, defaults will be used.

Usage

priors.spec(m0 = 0, CS0 = 3, n0 = 0.001, d0 = 0.001)

Arguments

m0	the value of the prior mean at time t=0, scalar (assumed to be the same for all nodes). The default is zero.
CS0	controls the scaling of the prior variance matrix C*_{0} at time t=0. The default is 3, giving a non-informative prior for C*_{0}, 3 x (p x p) identity matrix. p is the number of thetas.
n0	prior hyperparameter of precision phi ~ G(n_{0}/2; d_{0}/2). The default is a non-informative prior, with n0 = d0 = 0.001. n0 has to be higher than 0.
d0	prior hyperparameter of precision phi ~ G(n_{0}/2; d_{0}/2). The default is a non-informative prior, with n0 = d0 = 0.001.

Details

At time t=0, (theta_{0} | D_{0}, phi) ~ N(m_{0},C*_{0} x phi^{-1}), where D_{0} denotes the set of initial information.

Value

priors a list with the prior hyperparameters. Relevant to dlm.lpl, exhaustive.search, node, subject.

References

West, M. & Harrison, J., 1997. Bayesian Forecasting and Dynamic Models. Springer New York.

Examples

pr=priors.spec()
pr=priors.spec(n0=0.002)

DGM documentation built on Dec. 11, 2021, 9:54 a.m.