snow integration

How to use the <pkg>bnlearn</pkg> package with the parallel computing
  environment provided by the <pkg>snow</pkg> package.

documentation

Bayesian network structure learning (via constraint-based,
        score-based and hybrid algorithms), parameter learning (via ML
        and Bayesian estimators) and inference.  This package
        implements the Grow-Shrink (GS) algorithm, the Incremental
        Association (IAMB) algorithm, the Interleaved-IAMB (Inter-IAMB)
        algorithm, the Fast-IAMB (Fast-IAMB) algorithm, the Max-Min
        Parents and Children (MMPC) algorithm, the Hiton-PC algorithm,
        the ARACNE and Chow-Liu algorithms, the Hill-Climbing (HC)
        greedy search algorithm, the Tabu Search (TABU) algorithm, the
        Max-Min Hill-Climbing (MMHC) algorithm and the two-stage
        Restricted Maximization (RSMAX2) algorithm for both discrete
        and Gaussian networks, along with many score functions and
        conditional independence tests.  The Naive Bayes and the
        Tree-Augmented Naive Bayes (TAN) classifiers are also
        implemented.  Some utility functions (model comparison and
        manipulation, random data generation, arc orientation testing,
        simple and advanced plots) are included, as well as support for
        parameter estimation and inference, conditional probability
        queries and cross-validation.

Marco Scutari

bnlearn

Bayesian network structure learning, parameter learning and
        inference

snow integration function

# load snow, bnlearn and rsprng (for parallel random number
# generation, just in case it's needed); start LAM/MPI via
# lamboot if using an MPI cluster.
> library(snow)
> library(bnlearn)
> library(rsprng)
# initialize the cluster ("socket" and "PVM" clusters are fine, too).
> cl &lt;- makeCluster(2, type = "MPI")
Loading required package: Rmpi
        2 slaves are spawned successfully. 0 failed.
> clusterSetupSPRNG(cl)
# load the data.
> data(learning.test)
# call a learning function passing the cluster object (the
# return value of the previous makeCluster() call) as a
# parameter.
> res = gs(learning.test, cluster = cl)
# note that the number of test is evenly divided between
# the two nodes of the cluster.
> clusterEvalQ(cl, .test.counter)
[[1]]
[1] 47<p></p><p>[[2]]
[1] 42
# a few tests are still executed by this process.
> .test.counter
[1] 4
# stop the cluster.
> stopCluster(cl)
[1] 1</p>

snow integration: bnlearn - snow package integration

Description

Arguments

Parallel computing for constraint-based algorithms