public void ClassificationModelSelectingEnsembleLearner_Learn_Start_With_3_Models()
{
var learners = new IIndexedLearner <ProbabilityPrediction>[]
{
new ClassificationDecisionTreeLearner(2),
new ClassificationDecisionTreeLearner(5),
new ClassificationDecisionTreeLearner(7),
new ClassificationDecisionTreeLearner(9),
new ClassificationDecisionTreeLearner(11),
new ClassificationDecisionTreeLearner(21),
new ClassificationDecisionTreeLearner(23),
new ClassificationDecisionTreeLearner(1),
new ClassificationDecisionTreeLearner(14),
new ClassificationDecisionTreeLearner(17),
new ClassificationDecisionTreeLearner(19),
new ClassificationDecisionTreeLearner(33)
};
var metric = new LogLossClassificationProbabilityMetric();
var ensembleStrategy = new MeanProbabilityClassificationEnsembleStrategy();
var ensembleSelection = new ForwardSearchClassificationEnsembleSelection(metric, ensembleStrategy, 5, 3, true);
var sut = new ClassificationModelSelectingEnsembleLearner(learners, new RandomCrossValidation <ProbabilityPrediction>(5, 23),
ensembleStrategy, ensembleSelection);
var(observations, targets) = DataSetUtilities.LoadGlassDataSet();
var model = sut.Learn(observations, targets);
var predictions = model.PredictProbability(observations);
var actual = metric.Error(targets, predictions);
Assert.AreEqual(0.55183985816428427, actual, 0.0001);
}
public void ClassificationStackingEnsembleModel_PredictProbability_single()
{
var parser = new CsvParser(() => new StringReader(Resources.AptitudeData));
var observations = parser.EnumerateRows(v => v != "Pass").ToF64Matrix();
var targets = parser.EnumerateRows("Pass").ToF64Vector();
var rows = targets.Length;
var learners = new IIndexedLearner <ProbabilityPrediction>[]
{
new ClassificationDecisionTreeLearner(2),
new ClassificationDecisionTreeLearner(5),
new ClassificationDecisionTreeLearner(7),
new ClassificationDecisionTreeLearner(9)
};
var learner = new ClassificationStackingEnsembleLearner(learners, new ClassificationDecisionTreeLearner(9),
new RandomCrossValidation <ProbabilityPrediction>(5, 23), false);
var sut = learner.Learn(observations, targets);
var predictions = new ProbabilityPrediction[rows];
for (int i = 0; i < rows; i++)
{
predictions[i] = sut.PredictProbability(observations.Row(i));
}
var metric = new LogLossClassificationProbabilityMetric();
var actual = metric.Error(targets, predictions);
Assert.AreEqual(0.6696598716465223, actual, 0.0000001);
}
public void ClassificationBackwardEliminationModelSelectingEnsembleLearner_CreateMetaFeatures_Then_Learn()
{
var learners = new IIndexedLearner <ProbabilityPrediction>[]
{
new ClassificationDecisionTreeLearner(2),
new ClassificationDecisionTreeLearner(5),
new ClassificationDecisionTreeLearner(7),
new ClassificationDecisionTreeLearner(9),
new ClassificationDecisionTreeLearner(11),
new ClassificationDecisionTreeLearner(21),
new ClassificationDecisionTreeLearner(23),
new ClassificationDecisionTreeLearner(1),
new ClassificationDecisionTreeLearner(14),
new ClassificationDecisionTreeLearner(17),
new ClassificationDecisionTreeLearner(19),
new ClassificationDecisionTreeLearner(33)
};
var sut = new ClassificationBackwardEliminationModelSelectingEnsembleLearner(learners, 5);
var parser = new CsvParser(() => new StringReader(Resources.Glass));
var observations = parser.EnumerateRows(v => v != "Target").ToF64Matrix();
var targets = parser.EnumerateRows("Target").ToF64Vector();
var metaObservations = sut.LearnMetaFeatures(observations, targets);
var model = sut.SelectModels(observations, metaObservations, targets);
var predictions = model.PredictProbability(observations);
var metric = new LogLossClassificationProbabilityMetric();
var actual = metric.Error(targets, predictions);
Assert.AreEqual(0.52351727716455632, actual, 0.0001);
}
public void ClassificationEnsembleModel_PredictProbability_single()
{
var(observations, targets) = DataSetUtilities.LoadAptitudeDataSet();
var learners = new IIndexedLearner <ProbabilityPrediction>[]
{
new ClassificationDecisionTreeLearner(2),
new ClassificationDecisionTreeLearner(5),
new ClassificationDecisionTreeLearner(7),
new ClassificationDecisionTreeLearner(9)
};
var learner = new ClassificationEnsembleLearner(learners, new MeanProbabilityClassificationEnsembleStrategy());
var sut = learner.Learn(observations, targets);
var rows = targets.Length;
var predictions = new ProbabilityPrediction[rows];
for (int i = 0; i < rows; i++)
{
predictions[i] = sut.PredictProbability(observations.Row(i));
}
var metric = new LogLossClassificationProbabilityMetric();
var actual = metric.Error(targets, predictions);
Assert.AreEqual(0.32562112824941963, actual, 0.0000001);
}
public void ClassificationBackwardEliminationModelSelectingEnsembleLearner_CreateMetaFeatures_Then_Learn()
{
var learners = new IIndexedLearner <ProbabilityPrediction>[]
{
new ClassificationDecisionTreeLearner(2),
new ClassificationDecisionTreeLearner(5),
new ClassificationDecisionTreeLearner(7),
new ClassificationDecisionTreeLearner(9),
new ClassificationDecisionTreeLearner(11),
new ClassificationDecisionTreeLearner(21),
new ClassificationDecisionTreeLearner(23),
new ClassificationDecisionTreeLearner(1),
new ClassificationDecisionTreeLearner(14),
new ClassificationDecisionTreeLearner(17),
new ClassificationDecisionTreeLearner(19),
new ClassificationDecisionTreeLearner(33)
};
var sut = new ClassificationBackwardEliminationModelSelectingEnsembleLearner(learners, 5);
var(observations, targets) = DataSetUtilities.LoadGlassDataSet();
var metaObservations = sut.LearnMetaFeatures(observations, targets);
var model = sut.SelectModels(observations, metaObservations, targets);
var predictions = model.PredictProbability(observations);
var metric = new LogLossClassificationProbabilityMetric();
var actual = metric.Error(targets, predictions);
Assert.AreEqual(0.52351727716455632, actual, 0.0001);
}
public void ClassificationBackwardEliminationModelSelectingEnsembleLearner_Learn_Indexed()
{
var learners = new IIndexedLearner <ProbabilityPrediction>[]
{
new ClassificationDecisionTreeLearner(2),
new ClassificationDecisionTreeLearner(5),
new ClassificationDecisionTreeLearner(7),
new ClassificationDecisionTreeLearner(9),
new ClassificationDecisionTreeLearner(11),
new ClassificationDecisionTreeLearner(21),
new ClassificationDecisionTreeLearner(23),
new ClassificationDecisionTreeLearner(1),
new ClassificationDecisionTreeLearner(14),
new ClassificationDecisionTreeLearner(17),
new ClassificationDecisionTreeLearner(19),
new ClassificationDecisionTreeLearner(33)
};
var metric = new LogLossClassificationProbabilityMetric();
var ensembleStrategy = new MeanProbabilityClassificationEnsembleStrategy();
var sut = new ClassificationBackwardEliminationModelSelectingEnsembleLearner(learners, 5,
new RandomCrossValidation <ProbabilityPrediction>(5, 23), ensembleStrategy, metric);
var(observations, targets) = DataSetUtilities.LoadGlassDataSet();
var indices = Enumerable.Range(0, 25).ToArray();
var model = sut.Learn(observations, targets, indices);
var predictions = model.PredictProbability(observations);
var actual = metric.Error(targets, predictions);
Assert.AreEqual(2.3682546920482164, actual, 0.0001);
}
public void ClassificationForwardSearchModelSelectingEnsembleLearner_Learn()
{
var learners = new IIndexedLearner <ProbabilityPrediction>[]
{
new ClassificationDecisionTreeLearner(2),
new ClassificationDecisionTreeLearner(5),
new ClassificationDecisionTreeLearner(7),
new ClassificationDecisionTreeLearner(9),
new ClassificationDecisionTreeLearner(11),
new ClassificationDecisionTreeLearner(21),
new ClassificationDecisionTreeLearner(23),
new ClassificationDecisionTreeLearner(1),
new ClassificationDecisionTreeLearner(14),
new ClassificationDecisionTreeLearner(17),
new ClassificationDecisionTreeLearner(19),
new ClassificationDecisionTreeLearner(33)
};
var sut = new ClassificationForwardSearchModelSelectingEnsembleLearner(learners, 5);
var parser = new CsvParser(() => new StringReader(Resources.Glass));
var observations = parser.EnumerateRows(v => v != "Target").ToF64Matrix();
var targets = parser.EnumerateRows("Target").ToF64Vector();
var model = sut.Learn(observations, targets);
var predictions = model.PredictProbability(observations);
var metric = new LogLossClassificationProbabilityMetric();
var actual = metric.Error(targets, predictions);
Assert.AreEqual(0.51787562976713208, actual, 0.0001);
}
public void ClassificationForwardSearchModelSelectingEnsembleLearner_Learn_Start_With_3_Models()
{
var learners = new IIndexedLearner <ProbabilityPrediction>[]
{
new ClassificationDecisionTreeLearner(2),
new ClassificationDecisionTreeLearner(5),
new ClassificationDecisionTreeLearner(7),
new ClassificationDecisionTreeLearner(9),
new ClassificationDecisionTreeLearner(11),
new ClassificationDecisionTreeLearner(21),
new ClassificationDecisionTreeLearner(23),
new ClassificationDecisionTreeLearner(1),
new ClassificationDecisionTreeLearner(14),
new ClassificationDecisionTreeLearner(17),
new ClassificationDecisionTreeLearner(19),
new ClassificationDecisionTreeLearner(33)
};
var metric = new LogLossClassificationProbabilityMetric();
var ensembleStrategy = new MeanProbabilityClassificationEnsembleStrategy();
var sut = new ClassificationForwardSearchModelSelectingEnsembleLearner(learners, 5,
new StratifiedCrossValidation <ProbabilityPrediction>(5, 23), ensembleStrategy, metric, 3, true);
var parser = new CsvParser(() => new StringReader(Resources.Glass));
var observations = parser.EnumerateRows(v => v != "Target").ToF64Matrix();
var targets = parser.EnumerateRows("Target").ToF64Vector();
var model = sut.Learn(observations, targets);
var predictions = model.PredictProbability(observations);
var actual = metric.Error(targets, predictions);
Assert.AreEqual(0.54434276244488244, actual, 0.0001);
}
public void ClassificationRandomModelSelectingEnsembleLearner_Learn()
{
var learners = new IIndexedLearner <ProbabilityPrediction>[]
{
new ClassificationDecisionTreeLearner(2),
new ClassificationDecisionTreeLearner(5),
new ClassificationDecisionTreeLearner(7),
new ClassificationDecisionTreeLearner(9),
new ClassificationDecisionTreeLearner(11),
new ClassificationDecisionTreeLearner(21),
new ClassificationDecisionTreeLearner(23),
new ClassificationDecisionTreeLearner(1),
new ClassificationDecisionTreeLearner(14),
new ClassificationDecisionTreeLearner(17),
new ClassificationDecisionTreeLearner(19),
new ClassificationDecisionTreeLearner(33)
};
var sut = new ClassificationRandomModelSelectingEnsembleLearner(learners, 5);
var(observations, targets) = DataSetUtilities.LoadGlassDataSet();
var model = sut.Learn(observations, targets);
var predictions = model.PredictProbability(observations);
var metric = new LogLossClassificationProbabilityMetric();
var actual = metric.Error(targets, predictions);
Assert.AreEqual(0.60969181130388794, actual, 0.0001);
}
public void ClassificationRandomModelSelectingEnsembleLearner_Learn_Without_Replacement()
{
var learners = new IIndexedLearner <ProbabilityPrediction>[]
{
new ClassificationDecisionTreeLearner(2),
new ClassificationDecisionTreeLearner(5),
new ClassificationDecisionTreeLearner(7),
new ClassificationDecisionTreeLearner(9),
new ClassificationDecisionTreeLearner(11),
new ClassificationDecisionTreeLearner(21),
new ClassificationDecisionTreeLearner(23),
new ClassificationDecisionTreeLearner(1),
new ClassificationDecisionTreeLearner(14),
new ClassificationDecisionTreeLearner(17),
new ClassificationDecisionTreeLearner(19),
new ClassificationDecisionTreeLearner(33)
};
var metric = new LogLossClassificationProbabilityMetric();
var ensembleStrategy = new MeanProbabilityClassificationEnsembleStrategy();
var sut = new ClassificationRandomModelSelectingEnsembleLearner(learners, 5,
new StratifiedCrossValidation <ProbabilityPrediction>(5, 23), ensembleStrategy, metric, 1, false);
var(observations, targets) = DataSetUtilities.LoadGlassDataSet();
var model = sut.Learn(observations, targets);
var predictions = model.PredictProbability(observations);
var actual = metric.Error(targets, predictions);
Assert.AreEqual(0.5805783545646459, actual, 0.0001);
}
public void ClassificationStackingEnsembleModel_PredictProbability_single()
{
var(observations, targets) = DataSetUtilities.LoadAptitudeDataSet();
var learners = new IIndexedLearner <ProbabilityPrediction>[]
{
new ClassificationDecisionTreeLearner(2),
new ClassificationDecisionTreeLearner(5),
new ClassificationDecisionTreeLearner(7),
new ClassificationDecisionTreeLearner(9)
};
var learner = new ClassificationStackingEnsembleLearner(learners,
new ClassificationDecisionTreeLearner(9),
new RandomCrossValidation <ProbabilityPrediction>(5, 23), false);
var sut = learner.Learn(observations, targets);
var rows = targets.Length;
var predictions = new ProbabilityPrediction[rows];
for (int i = 0; i < rows; i++)
{
predictions[i] = sut.PredictProbability(observations.Row(i));
}
var metric = new LogLossClassificationProbabilityMetric();
var actual = metric.Error(targets, predictions);
Assert.AreEqual(0.6696598716465223, actual, 0.0000001);
}
public void ClassificationEnsembleModel_PredictProbability_Multiple()
{
var parser = new CsvParser(() => new StringReader(Resources.AptitudeData));
var observations = parser.EnumerateRows(v => v != "Pass").ToF64Matrix();
var targets = parser.EnumerateRows("Pass").ToF64Vector();
var rows = targets.Length;
var learners = new IIndexedLearner <ProbabilityPrediction>[]
{
new ClassificationDecisionTreeLearner(2),
new ClassificationDecisionTreeLearner(5),
new ClassificationDecisionTreeLearner(7),
new ClassificationDecisionTreeLearner(9)
};
var learner = new ClassificationEnsembleLearner(learners, new MeanProbabilityClassificationEnsembleStrategy());
var sut = learner.Learn(observations, targets);
var predictions = sut.PredictProbability(observations);
var metric = new LogLossClassificationProbabilityMetric();
var actual = metric.Error(targets, predictions);
Assert.AreEqual(0.32562112824941963, actual, 0.0000001);
}
public void CrossValidation_CrossValidate_ProbabilityPredictions()
{
#region Read data
// Use StreamReader(filepath) when running from filesystem
var parser = new CsvParser(() => new StringReader(Resources.winequality_white));
var targetName = "quality";
// read feature matrix (all columns different from the targetName)
var observations = parser.EnumerateRows(c => c != targetName)
.ToF64Matrix();
// read targets
var targets = parser.EnumerateRows(targetName)
.ToF64Vector();
#endregion
// creates cross validator, observations are shuffled randomly
var cv = new RandomCrossValidation <ProbabilityPrediction>(crossValidationFolds: 5, seed: 42);
// create learner
var learner = new ClassificationDecisionTreeLearner();
// cross-validated predictions
var cvPredictions = cv.CrossValidate(learner, observations, targets);
// metric for measuring model error
var metric = new LogLossClassificationProbabilityMetric();
// cross-validation provides an estimate on how the model will perform on unseen data
Trace.WriteLine("Cross-validation error: " + metric.Error(targets, cvPredictions));
// train and predict training set for comparison
var predictions = learner.Learn(observations, targets).PredictProbability(observations);
// The training set is NOT a good estimate of how well the model will perfrom on unseen data.
Trace.WriteLine("Training error: " + metric.Error(targets, predictions));
}
public void LogLossClassificationMetric_Error_2()
{
var sut = new LogLossClassificationProbabilityMetric(1e-15);
var predictions = new ProbabilityPrediction[] {
new ProbabilityPrediction(0, new Dictionary <double, double> {
{ 0, 1.0 }, { 1, 1.0 }, { 2, 1.0 }
}),
new ProbabilityPrediction(1, new Dictionary <double, double> {
{ 0, 0.0 }, { 1, 1.0 }, { 2, 0.0 }
}),
new ProbabilityPrediction(2, new Dictionary <double, double> {
{ 0, 0.0 }, { 1, 0.0 }, { 2, 1.0 }
}),
};
var targets = new double[] { 0, 1, 2 };
var actual = sut.Error(targets, predictions);
Assert.AreEqual(0.36620409622270467, actual, 0.0001);
}
public void LogLossClassificationMetric_Error_1()
{
var sut = new LogLossClassificationProbabilityMetric(1e-15);
var predictions = new ProbabilityPrediction[] {
new ProbabilityPrediction(0, new Dictionary <double, double> {
{ 0, 1.0 }, { 1, 0.0 }, { 2, 0.0 }
}),
new ProbabilityPrediction(1, new Dictionary <double, double> {
{ 0, 0.0 }, { 1, 1.0 }, { 2, 0.0 }
}),
new ProbabilityPrediction(2, new Dictionary <double, double> {
{ 0, 0.0 }, { 1, 0.0 }, { 2, 1.0 }
}),
};
var targets = new double[] { 0, 1, 2 };
var actual = sut.Error(targets, predictions);
Assert.AreEqual(9.9920072216264128e-16, actual, 1e-17);
}
public void ClassificationModelSelectingEnsembleLearner_Learn_Indexed()
{
var learners = new IIndexedLearner <ProbabilityPrediction>[]
{
new ClassificationDecisionTreeLearner(2),
new ClassificationDecisionTreeLearner(5),
new ClassificationDecisionTreeLearner(7),
new ClassificationDecisionTreeLearner(9),
new ClassificationDecisionTreeLearner(11),
new ClassificationDecisionTreeLearner(21),
new ClassificationDecisionTreeLearner(23),
new ClassificationDecisionTreeLearner(1),
new ClassificationDecisionTreeLearner(14),
new ClassificationDecisionTreeLearner(17),
new ClassificationDecisionTreeLearner(19),
new ClassificationDecisionTreeLearner(33)
};
var metric = new LogLossClassificationProbabilityMetric();
var ensembleStrategy = new MeanProbabilityClassificationEnsembleStrategy();
var ensembleSelection = new ForwardSearchClassificationEnsembleSelection(metric, ensembleStrategy, 5, 1, true);
var sut = new ClassificationModelSelectingEnsembleLearner(learners, new RandomCrossValidation <ProbabilityPrediction>(5, 23),
ensembleStrategy, ensembleSelection);
var parser = new CsvParser(() => new StringReader(Resources.Glass));
var observations = parser.EnumerateRows(v => v != "Target").ToF64Matrix();
var targets = parser.EnumerateRows("Target").ToF64Vector();
var indices = Enumerable.Range(0, 25).ToArray();
var model = sut.Learn(observations, targets, indices);
var predictions = model.PredictProbability(observations);
var actual = metric.Error(targets, predictions);
Assert.AreEqual(2.3682546920482164, actual, 0.0001);
}
