Learning Bayesian Networks with the bnlearn R Package文档.docVIP

JournalofStatisticalSoftware July2010,Volume35,Issue3. / Learning Bayesian Networks with the bnlearn R Package Marco Scutari UniversityofPadova Abstract bnlearnisanRpackage(RDevelopmentCoreTeam2010)whichincludesseveralalgo- rithmsforlearningthestructureofBayesiannetworkswitheitherdiscreteorcontinuous variables. Both constraint-based and score-based algorithms are implemented, and can usethefunctionalityprovidedbythesnowpackage(Tierneyetal.2008)toimprovetheir performanceviaparallelcomputing. Severalnetworkscoresandconditionalindependence algorithmsareavailableforboththelearningalgorithmsandindependentuse. Advanced plottingoptionsareprovidedbytheRgraphvizpackage(Gentryetal.2010). Keywords:bayesiannetworks,R,structurelearningalgorithms,constraint-basedalgorithms, score-basedalgorithms,conditionalindependencetests. 1. Introduction In recent years Bayesian networks have been used in many ?elds, from On-line Analytical Processing (OLAP) performance enhancement (Margaritis 2003) to medical service perfor- mance analysis (Acid et al. 2004), gene expression analysis (Friedman et al. 2000), breast cancerprognosisandepidemiology(HolmesandJain2008). Thehighdimensionalityofthedatasetscommoninthesedomainshaveledtothedevelop- mentofseverallearningalgorithmsfocusedonreducingcomputationalcomplexitywhilestill learning the correct network. Some examples are the grow-shrink algorithm in Margaritis (2003),theincrementalassociationalgorithmanditsderivativesinTsamardinosetal.(2003) andinYaramakalaandMargaritis(2005),thesparsecandidate algorithminFriedmanetal. (1999),theoptimalreinsertioninMooreandWong(2003)andthegreedyequivalentsearch inChickering(2002). The aim of the bnlearn package is to provide a free implementation of some of these struc- ture learning algorithms along with the conditional independence tests and network scores  2 LearningBayesianNetworkswiththebnlearnRPackage used to construct the Bayesian network. Both discrete and continuous data are supported. Fu