public void RandomRegressionEnsembleSelection_Constructor_Number_Of_Availible_Models_Lower_Than_Number_Of_Models_To_Select() { var sut = new RandomRegressionEnsembleSelection(new MeanSquaredErrorRegressionMetric(), new MeanRegressionEnsembleStrategy(), 5, 1, true); var observations = new F64Matrix(10, 3); var targets = new double[10]; sut.Select(observations, targets); }
public void RandomRegressionEnsembleSelection_Select_No_Replacements() { var sut = new RandomRegressionEnsembleSelection(new MeanSquaredErrorRegressionMetric(), new MeanRegressionEnsembleStrategy(), 3, 10, false); var random = new Random(42); var observations = new F64Matrix(10, 10); observations.Map(() => random.Next()); var targets = Enumerable.Range(0, 10).Select(v => random.NextDouble()).ToArray(); var actual = sut.Select(observations, targets); var expected = new int[3] { 3, 8, 2 }; CollectionAssert.AreEqual(expected, actual); }
public void RandomRegressionEnsembleSelection_Constructor_EnsembleStratey_Null() { var sut = new RandomRegressionEnsembleSelection(new MeanSquaredErrorRegressionMetric(), null, 5, 1, true); }
public void RandomRegressionEnsembleSelection_Constructor_Number_Of_Models_From_Start_Too_Low() { var sut = new RandomRegressionEnsembleSelection(new MeanSquaredErrorRegressionMetric(), new MeanRegressionEnsembleStrategy(), 5, 0, true); }
public void RandomRegressionEnsembleSelection_Constructor_Metric_Null() { var sut = new RandomRegressionEnsembleSelection(null, new MeanRegressionEnsembleStrategy(), 5, 1, true); }