make_structure_move {EDISON}

Makes a structure move.

Description

This function makes a network structure move.

Usage

make_structure_move(x, y, S, B, Sig2, q, qmax, network.info, method, Mphase, E,
  fixed.edges, HYPERvar)

Arguments

x

Response data.

y

Target data.

S

Network structure for the current target node, a NumSegs by NumNodes matrix.

B

Same as S, but including the regression parameters.

Sig2

Sigma squared parameters.

q

Number of nodes.

qmax

Maximum number of parents.

network.info

Network information, as collected by CollectNetworkInfo.

method

Information sharing method: Either 'poisson', 'exp_hard',

'exp_soft', 'bino_hard', 'bino_soft'.

Mphase

Segment boundary positions.

E

Changepoint vector.

fixed.edges

Matrix of size NumNodes by NumNodes, with fixed.edges[i,j]==1|0 if the edge between nodes i and j is fixed, and -1 otherwise. Defaults to NULL (no edges fixed).

HYPERvar

Hyperparameter variables.

Value

Returns a list containing the following elements:

newS

Updated network structure.

newB

Updated network structure with regression parameters.

move

Type of move being made: 1 for network structure moves.

accept

1 if the move has been accepted, 0 otherwise.

Author(s)

Frank Dondelinger

References

For more information about the MCMC moves, see:

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