private void LearnNewModel()
{
previousModel = new List <Feature>(currentModel);
GlobalState.logInfo.logLine("Learning set: " + currentLearningSet.Count + ", Validation set: "
+ currentLearningSet.Count);
Learning exp = new Learning(currentLearningSet, currentValidationSet)
{
metaModel = influenceModel,
mlSettings = this.mlSettings
};
exp.learn(currentModel);
if (exp.models.Count != 1)
{
GlobalState.logError.logLine("There should be exactly one learned model! Aborting active learning!");
Environment.Exit(0);
}
FeatureSubsetSelection fss = exp.models[0];
currentModel = fss.LearningHistory.Last().FeatureSet;
previousValidationError = currentValidationError;
currentValidationError = fss.finalError;
currentGlobalError = fss.computeError(currentModel, GlobalState.allMeasurements.Configurations, false);
GlobalState.logInfo.logLine("globalError = " + currentGlobalError);
}
/// <summary>
/// Continues learning with recovered learning data.
/// </summary>
/// <param name="recoveredLr">Learning rounds that were already performed.</param>
public void continueLearning(List <LearningRound> recoveredLr)
{
if (!hasNecessaryData())
{
return;
}
if (this.mlSettings.bagging)
{
throw new NotImplementedException("Recovering with bagging currently dosent work");
}
else
{
ObservableCollection <LearningRound> learningRounds = new ObservableCollection <LearningRound>();
GlobalState.logInfo.logLine("Learning progress:");
foreach (LearningRound lr in recoveredLr)
{
GlobalState.logInfo.logLine(lr.ToString());
learningRounds.Add(lr);
}
InfluenceModel infMod = new InfluenceModel(GlobalState.varModel, GlobalState.currentNFP);
FeatureSubsetSelection sel = new FeatureSubsetSelection(infMod, this.mlSettings);
this.models.Add(sel);
sel.setLearningSet(testSet);
sel.setValidationSet(this.validationSet);
Stopwatch sw = new Stopwatch();
sw.Start();
sel.continueLearn(learningRounds);
sw.Stop();
Console.WriteLine("Elapsed={0}", sw.Elapsed);
}
}
public void learn()
{
if (!hasNecessaryData())
{
return;
}
if (this.mlSettings.bagging)
{
//Get number of cores
int coreCount = System.Environment.ProcessorCount;
createThreadPool(coreCount);
this.nbBaggings = this.mlSettings.baggingNumbers;
iCount = this.nbBaggings;
Random rand = new Random(0);
int nbOfConfigs = (testSet.Count * this.mlSettings.baggingTestDataFraction) / 100;
for (int i = 0; i < nbBaggings; i++)
{
InfluenceModel infMod = new InfluenceModel(GlobalState.varModel, GlobalState.currentNFP);
FeatureSubsetSelection sel = new FeatureSubsetSelection(infMod, this.mlSettings);
this.models.Add(sel);
List <int> selection = new List <int>();
for (int r = 0; r <= nbOfConfigs; r++)
{
selection.Add(rand.Next(nbOfConfigs));
}
List <Configuration> newTestSet = new List <Configuration>();
List <Configuration> newValidationSet = new List <Configuration>();
for (int r = 0; r <= selection.Count; r++)
{
if (selection.Contains(r))
{
newTestSet.Add(testSet[r]);
}
else
{
newValidationSet.Add(testSet[r]);
}
}
sel.setLearningSet(newTestSet);
sel.setValidationSet(newValidationSet);
Task task = EnqueueTask(() => sel.learn());
}
eventX.WaitOne(Timeout.Infinite, true);
averageModels();
}
else
{
GlobalState.logInfo.logLine("Learning progress:");
InfluenceModel infMod = new InfluenceModel(GlobalState.varModel, GlobalState.currentNFP);
FeatureSubsetSelection sel = new FeatureSubsetSelection(infMod, this.mlSettings);
this.models.Add(sel);
sel.setLearningSet(testSet);
sel.setValidationSet(this.validationSet);
Stopwatch sw = new Stopwatch();
sw.Start();
sel.learn();
sw.Stop();
Console.WriteLine("Elapsed={0}", sw.Elapsed);
}
}
