EDISON-package {EDISON} | R Documentation |
This package runs an MCMC simulation to reconstruct networks from time series data, using a non-homogeneous, time-varying dynamic Bayesian network. Networks segments and changepoints are inferred concurrently, and information sharing priors provide a reduction of the inference uncertainty.
Package: | EDISON |
Type: | Package |
Version: | 1.1.1 |
Date: | 2016-03-30 |
License: | GPL-2 |
LazyLoad: | yes |
Frank Dondelinger, Sophie Lebre
Maintainer: Frank Dondelinger <fdondelinger.work@gmail.com>
Dondelinger et al. (2012), "Non-homogeneous dynamic Bayesian networks with Bayesian regularization for inferring gene regulatory networks with gradually time-varying structure", Machine Learning.
Husmeier et al. (2010), "Inter-time segment information sharing for non-homogeneous dynamic Bayesian networks", NIPS.
# Generate random gene network and simulate data from it
dataset = simulateNetwork(l=25)
# Run MCMC simulation to infer networks and changepoint locations
result = EDISON.run(dataset$sim_data, num.iter=500)
# Calculate posterior probabilities of changepoints
cps = calculateCPProbabilities(result)
# Calculate marginal posterior probabilities of edges in the network
network = calculateEdgeProbabilities(result)