#Getting Started
##Build a pool of M original dataset samples, each of size N (sample with replacement)
dataPool = getDataPerturbationPool(dataset, N, M, parallelCluster)##Compute partially weighted graphs pool with random edges sampling, with an edges sampling ratio of E
fullEL = getAdmissibleEdgeSpace(dataset, TRUE, parentIds)
graphsPool = getWeightedGraphsPool(dataset, dataPool, E, fullEL, symmetricEdgeSampling, logThresholdPre, score, parentIds, parallelCluster),##Learn arborescence model components pool from these graphs
components = learnComponentsPool(graphsPool, fullEL, toMaximize, learnFun = Learning.arborescence.fromWeightedGraph, reverse = FALSE, parallelCluster = parallelCluster)##Post processing components via pruning edges with unsignificant scores (here, the ones under 0)
components = pruningComponentsFunction(components, 0, TRUE)##Combine the M components
compositeModel = combineComponentsInPool(components, M, 1, function(count){count / m})##Learn both best and worst component sets, before aggregating
topComponents = learnComponentsPool(graphsPool, fullEL, toMaximize, learnFun = Learning.arborescence.fromWeightedGraph, reverse = FALSE, parallelCluster = parallelCluster)
topComponents = pruningComponentsFunction(topComponents, 0, toMaximize)
bottomComponents = learnComponentsPool(graphsPool, fullEL, !toMaximize, learnFun = Learning.arborescence.fromWeightedGraph, reverse = FALSE, parallelCluster = parallelCluster)
bottomComponents = pruningComponentsFunction(bottomComponents, 0, !toMaximize)
mergedComponents = pairComponentPools(topComponents, bottomComponents, ComponentsAggregator.concat)
compositeModel = combineComponentsInPool(mergedComponents, M, 1, function(count){count / m})