EDISON (version 1.1.1)

ExpHyperMove: Makes an exponential hyperparameter move.

Description

This function tries to make a level-1 or level-2 hyperparameter move for the exponential prior

Usage

ExpHyperMove(network.info, node.sharing, GLOBvar, hyper.proposals)

Arguments

network.info

The network information collected by CollectNetworkInfo.

node.sharing

The type of information sharing among nodes: 'soft' or 'hard'.

GLOBvar

Collection of global variables of the MCMC.

hyper.proposals

Proposal width of the hyperparameter move.

Value

Returns a list with elements:

move.made

1 if a level-1 hyperparameter move has been made, 0 otherwise.

network.info

Network information with updated hyperparameters if the move was accepted.

accept

Whether a level-1 hyperparameter move has been accepted or not.

move.made.k

1 if a level-2 hyperparameter move has been made, 0 otherwise.

accept.k

Whether a level-2 hyperparameter move has been accepted or not.

move

Type of move: 2 for a level-1 hyperparameter move, 3 for a level-2 hyperparameter move.

References

For information about the exponential information sharing prior, see:

Husmeier et al. (2010), "Inter-time segment information sharing for non-homogeneous dynamic Bayesian networks", NIPS.

Dondelinger et al. (2012), "Non-homogeneous dynamic Bayesian networks with Bayesian regularization for inferring gene regulatory networks with gradually time-varying structure", Machine Learning.

See Also

ExpHyperRatioTarget