public override ConfusionMatrix Execute()
{
//Create an network with one layer and one neuron in that layer
var network = new ActivationNetwork(new ThresholdFunction(), 3, 1);
//Bind the reference of the neuron
var neuron = network.Layers[0].Neurons[0] as ActivationNeuron;
//Create the Perceptron learning algorithm
//Library perceptron implements a single layer linear classifier
var teacher = new PerceptronLearning(network);
teacher.LearningRate = 0.1;
//Enrich the dimensions of the vectors, padding 1 to the end
var richTraining = AlgorithmHelpers.PaddDimension(trainingSet);
var richTesting = AlgorithmHelpers.PaddDimension(testSet);
//Training the network until the error is small enough
//or 500 hundred iterations have been computed
int epochs = 0;
while (true)
{
double error = teacher.RunEpoch(richTraining, trainingOutput);/// trainingSet.Length;
++epochs;
if (error < 0.025 * trainingSet.Length || epochs == 500) break;
}
var predicted = richTesting
.Select(x => neuron.Compute(x))
.Select(x => Convert.ToInt32(x))
.ToArray();
//Create a confusion matrix with the calculated parameters
ConfusionMatrix cmatrix = new ConfusionMatrix(predicted, expected, POSITIVE, NEGATIVE);
OnAlgorithmEnded(Enumerable.Repeat(neuron, 1), cmatrix);
return cmatrix;
}
public void FisherExactTestConstructorTest1()
{
// Example from http://rfd.uoregon.edu/files/rfd/StatisticalResources/lec_05a.txt
ConfusionMatrix matrix = new ConfusionMatrix
(
14, 10,
21, 3
);
{
var target = new FisherExactTest(matrix, OneSampleHypothesis.ValueIsSmallerThanHypothesis);
Assert.AreEqual(OneSampleHypothesis.ValueIsSmallerThanHypothesis, target.Hypothesis);
Assert.AreEqual(DistributionTail.OneLower, target.Tail);
Assert.AreEqual(0.02450, target.PValue, 1e-5);
}
{
var target = new FisherExactTest(matrix, OneSampleHypothesis.ValueIsDifferentFromHypothesis);
Assert.AreEqual(OneSampleHypothesis.ValueIsDifferentFromHypothesis, target.Hypothesis);
Assert.AreEqual(DistributionTail.TwoTail, target.Tail);
Assert.AreEqual(0.04899, target.PValue, 1e-4);
}
{
var target = new FisherExactTest(matrix, OneSampleHypothesis.ValueIsGreaterThanHypothesis);
Assert.AreEqual(OneSampleHypothesis.ValueIsGreaterThanHypothesis, target.Hypothesis);
Assert.AreEqual(DistributionTail.OneUpper, target.Tail);
Assert.AreEqual(0.99607, target.PValue, 1e-4);
}
}
public virtual List<ConfusionMatrix> TestModel(TrainningData trainningData)
{
int[] expected = trainningData.ClassificationAttribute;
int[] predicted = ComputeModel(trainningData.TrainningAttributes);
int positiveValue = trainningData.PositiveValue;
int negativeValue = trainningData.NegativeValue;
ConfusionMatrix confusionMatrix = new ConfusionMatrix(predicted, expected, positiveValue, negativeValue);
return new List<ConfusionMatrix> { confusionMatrix };
}
public override ConfusionMatrix Execute()
{
var richTesting = testSet;
ConcurrentDictionary<double, ActivationNetwork> networks = new ConcurrentDictionary<double, ActivationNetwork>();
Parallel.For(0, 2000, (int i) =>
{
//Create an activation network
ThreadLocal<ActivationNetwork> network = new ThreadLocal<ActivationNetwork>(() =>
{
return new ActivationNetwork(new SigmoidFunction(2), 2, 2, 1);
});
ThreadLocal<ResilientBackpropagationLearning> teachear = new ThreadLocal<ResilientBackpropagationLearning>(() =>
{
return new ResilientBackpropagationLearning(network.Value);
});
ThreadLocal<int> iter = new ThreadLocal<int>(() => { return 0; });
ThreadLocal<double> error = new ThreadLocal<double>(() => { return 0; });
while (networks.IsEmpty)
{
error.Value = teachear.Value.RunEpoch(trainingSet, trainingOutput);
iter.Value++;
if (iter.Value == 1000) break;
}
if (!networks.ContainsKey(error.Value)) networks.TryAdd(error.Value, network.Value);
}
);
Dictionary<ConfusionMatrix, ActivationNetwork> cms = new Dictionary<ConfusionMatrix, ActivationNetwork>();
foreach (var keyv in networks)
{
var p = richTesting
.Select(x => keyv.Value.Compute(x))
.Select(x => Convert.ToInt32(x[0]))
.ToArray();
ConfusionMatrix cm = new ConfusionMatrix(p, expected, POSITIVE, NEGATIVE);
cms.Add(cm, keyv.Value);
}
var kv = (from x in cms
orderby x.Key.Accuracy descending
select x).First();
var neurons = from neuron in kv.Value.Layers[0].Neurons
select neuron as ActivationNeuron;
OnAlgorithmEnded(neurons, kv.Key);
return kv.Key;
}
public void OnAlgorithmEnded(IEnumerable<ActivationNeuron> neurons, ConfusionMatrix matrix)
{
var handler = Finished;
if (handler != null)
{
var e = new AlgorithmFinishedEventArgs()
{
Neurons = neurons,
Matrix = matrix
};
handler(this, e);
}
}
/// <summary>
/// Creates a new McNemar test.
/// </summary>
///
/// <param name="matrix">The contingency table to test.</param>
/// <param name="yatesCorrection">True to use Yate's correction of
/// continuity, falser otherwise. Default is false.</param>
///
public McNemarTest(ConfusionMatrix matrix, bool yatesCorrection = false)
{
int a = matrix.TruePositives;
int b = matrix.FalseNegatives;
int c = matrix.FalsePositives;
int d = matrix.TrueNegatives;
double u = (yatesCorrection) ? Math.Abs(b - c) - 0.5 : u = b - c;
double chiSquare = (u * u) / (b + c);
int df = 1;
Compute(chiSquare, df);
}
public void ConfusionMatrixConstructorTest()
{
// The correct and expected output values (as confirmed by a Gold
// standard rule, actual experiment or true verification)
int[] expected = { 0, 0, 1, 0, 1, 0, 0, 0, 0, 0 };
// The values as predicted by the decision system or
// the test whose performance is being measured.
int[] predicted = { 0, 0, 0, 1, 1, 0, 0, 0, 0, 1 };
// In this test, 1 means positive, 0 means negative
int positiveValue = 1;
int negativeValue = 0;
// Create a new confusion matrix using the given parameters
ConfusionMatrix matrix = new ConfusionMatrix(predicted, expected,
positiveValue, negativeValue);
// At this point,
// True Positives should be equal to 1;
// True Negatives should be equal to 6;
// False Negatives should be equal to 1;
// False Positives should be equal to 2.
int falseNegatives = 1;
int falsePositives = 2;
int truePositives = 1;
int trueNegatives = 6;
Assert.AreEqual(predicted.Length, matrix.Samples);
Assert.AreEqual(8, matrix.ActualNegatives);
Assert.AreEqual(2, matrix.ActualPositives);
Assert.AreEqual(7, matrix.PredictedNegatives);
Assert.AreEqual(3, matrix.PredictedPositives);
Assert.AreEqual(falseNegatives, matrix.FalseNegatives);
Assert.AreEqual(falsePositives, matrix.FalsePositives);
Assert.AreEqual(truePositives, matrix.TruePositives);
Assert.AreEqual(trueNegatives, matrix.TrueNegatives);
Assert.AreEqual(0.7, matrix.Accuracy);
Assert.AreEqual(0.5, matrix.Sensitivity);
Assert.AreEqual(0.75, matrix.Specificity);
Assert.AreEqual((0.5 + 0.75) / 2.0, matrix.Efficiency);
Assert.AreEqual(0.21821789023599239, matrix.MatthewsCorrelationCoefficient);
}
public override List<ConfusionMatrix> TestModel(TrainningData trainningData)
{
ContinuousDataTableAdapter continuousDataTableAdapter = new ContinuousDataTableAdapter();
DataTable continuousDataTable = continuousDataTableAdapter.GetData();
DataTable dataTable = continuousDataTable.DefaultView.ToTable(false, TableMetaData.TestingAttributes);
string[] columnNames;
double[][] inputs = dataTable.ToArray(out columnNames);
int[] expected = trainningData.ClassificationAttribute;
int[] predicted = ComputeModel(inputs);
int positiveValue = trainningData.PositiveValue;
int negativeValue = trainningData.NegativeValue;
ConfusionMatrix confusionMatrix = new ConfusionMatrix(predicted, expected, positiveValue, negativeValue);
return new List<ConfusionMatrix> { confusionMatrix };
}
/// <summary>
/// Validate than Accord.Net DecisionTree or MulticlassSupportVectorMachine class
/// </summary>
/// <param name="test">The test dataset</param>
/// <param name="obj">The Accord.Net DecisionTree or MulticlassSupportVectorMachine object</param>
public static void Validate(DataTable test, object obj)
{
List<int> expected = new List<int>();
List<int> predicted = new List<int>();
foreach (DataRow row in test.Rows)
{
double gender = 1;
if (string.Compare((string)row["Gender"], "F", true, CultureInfo.CurrentCulture) == 0)
{
gender = 0;
}
double[] testQuery = new double[]
{
gender, Convert.ToDouble(row["YearOfBirth"], CultureInfo.CurrentCulture), Convert.ToDouble(row["SmokingEffectiveYear"], CultureInfo.CurrentCulture), Convert.ToDouble(row["NISTcode"], CultureInfo.CurrentCulture),
Convert.ToDouble(row["Height"], CultureInfo.CurrentCulture), Convert.ToDouble(row["Weight"], CultureInfo.CurrentCulture), Convert.ToDouble(row["BMI"], CultureInfo.CurrentCulture),
Convert.ToDouble(row["SystolicBP"], CultureInfo.CurrentCulture), Convert.ToDouble(row["DiastolicBP"], CultureInfo.CurrentCulture), Convert.ToDouble(row["RespiratoryRate"], CultureInfo.CurrentCulture),
Convert.ToDouble(row["Temperature"], CultureInfo.CurrentCulture)
};
int output = -1;
if (obj is DecisionTree)
{
output = ((DecisionTree)obj).Compute(testQuery);
}
else if (obj is MulticlassSupportVectorMachine)
{
output = ((MulticlassSupportVectorMachine)obj).Compute(testQuery);
}
else
{
throw new ArgumentException("Unknown algorithm for validation.");
}
expected.Add(Convert.ToInt32(row["DMIndicator"], CultureInfo.CurrentCulture));
predicted.Add(output);
}
var confusionMatrix = new ConfusionMatrix(predicted.ToArray(), expected.ToArray());
Logger.Info("The following is the confusion matrix (aka truth table). Look for TP (true positive), FP (false positive), etc...");
Logger.Info("Accuracy :" + confusionMatrix.ToString());
Logger.Info("Hit Enter to continue....");
Console.ReadLine();
}
public override ConfusionMatrix Execute()
{
//Create a knn classifer with 2 classes
var knn = new KNearestNeighbors(k: k,
classes: 2,
inputs: trainingSet,
outputs: trainingOutput);
//Map the classifier over the test set
//This wil return an array where index i is the classificatioon of the i-th vector
//of the testSet
var predicted = AlgorithmHelpers
.MergeArrays(trainingSet, testSet)
.Select(x => knn.Compute(x))
.ToArray();
//Create a new confusion matrix with the calculated parameters
var cmatrix = new ConfusionMatrix(predicted, AlgorithmHelpers.MergeArrays(trainingOutput, expected), POSITIVE, NEGATIVE);
return cmatrix;
}
public void McNemarTestConstructorTest()
{
int[,] matrix =
{
{ 101, 121 },
{ 59, 33 },
};
ConfusionMatrix a = new ConfusionMatrix(matrix);
McNemarTest target = new McNemarTest(a, true);
Assert.AreEqual(21.0125, target.Statistic);
Assert.AreEqual(1, target.DegreesOfFreedom);
McNemarTest target2 = new McNemarTest(a, false);
Assert.AreEqual(21.355555, target2.Statistic, 1e-5);
Assert.AreEqual(1, target2.DegreesOfFreedom);
}
public override ConfusionMatrix Execute()
{
//The Least Squares algorithm
//It uses a PartialLeastSquaresAnalysis library object using a non-linear iterative partial least squares algorithm
//and runs on the mean-centered and standardized data
//Create an analysis
var pls = new PartialLeastSquaresAnalysis(trainingSet,
trainingOutput,
AnalysisMethod.Standardize,
PartialLeastSquaresAlgorithm.NIPALS);
pls.Compute();
//After computing the analysis
//create a linear model to predict new variables
MultivariateLinearRegression regression = pls.CreateRegression();
//This will hold the result of the classifications
var predictedLifted = new int[testSet.GetLength(0)][];
for (int i = 0; i < predictedLifted.Length; ++i)
{
predictedLifted[i] = regression
.Compute(testSet.GetRow(i)) //Retrieve the row vector of the test set
.Select(x => Convert.ToInt32(x))// Convert the result to int
.ToArray();
}
//Unlift the prediction vector
var predicted = predictedLifted
.SelectMany(x => x)
.ToArray();
//Create a new confusion matrix with the calculated parameters
ConfusionMatrix cmatrix = new ConfusionMatrix(predicted, expected, POSITIVE, NEGATIVE);
return cmatrix;
}
//This is the function that runs the demo application
//It handles the file loading the execution of the algorithms
//and the manipulation of the controls
private void SearchSolution()
{
//Read the data from the files
Double[][] file1DataRaw = UtilityProvider.ReadMatrixFromFile(@"class_1.dat");
Double[][] file2DataRaw = UtilityProvider.ReadMatrixFromFile(@"class_2.dat");
//Chose 2 features
class_1 = UtilityProvider.ScaleDown(UtilityProvider.ChooseFeatures(file1DataRaw));
class_2 = UtilityProvider.ScaleDown(UtilityProvider.ChooseFeatures(file2DataRaw));
//Fill the charts with the data
ShowTrainingData();
//Clear the list view
ClearListView();
//Fill it with the confusion matrix for each algorithm per iteration
ConfusionMatrix[,] statistics = new ConfusionMatrix[4, 5];
//Merge the two data sets
//and run kmeans
RunKMeans(UtilityProvider.MergeArrays(file1DataRaw, file2DataRaw));
this.Invoke(new Action(() => progressBar1.Step *= 2));
//Partition 5 times and run the algorithms
for (int i = 0; i < 5; ++i)
{
this.Invoke(new Action(() => progressBar1.PerformStep()));
//Partition its class to training and testing set
var partitions = new DataPartition[] { UtilityProvider.Partition(class_1), UtilityProvider.Partition(class_2) };
//Create the training data
var trainingPair = UtilityProvider.CreateDataPair(partitions[0].Item1, partitions[1].Item1);
var trainingSet = trainingPair.Item1;
var trainingOutput = trainingPair.Item2;
//Create the testing data
var testingPair = UtilityProvider.CreateDataPair(partitions[0].Item2, partitions[1].Item2);
var testingSet = testingPair.Item1;
var testingOutput = testingPair.Item2;
//Some functions need the training output to be a vector of doubles
var doubleTO = trainingOutput
.Select(x =>
new Double[] { Convert.ToDouble(x) })
.ToArray();
for (int k = 1; k < 3; ++k)
{
if (BestKNN == null)
{
BestKNN = RunKNN(k, trainingSet, trainingOutput, testingSet, testingOutput);
}
else
{
var iter = RunKNN(k, trainingSet, trainingOutput, testingSet, testingOutput);
if (iter.Accuracy > BestKNN.Accuracy)
BestKNN = iter;
}
}
//Compute the confusion matrices for the four classifiers
statistics[0, i] = RunPerceptron(trainingSet, doubleTO, testingSet, testingOutput);
statistics[1, i] = RunLS(UtilityProvider.JaggedToMD(trainingSet), UtilityProvider.JaggedToMD(doubleTO), UtilityProvider.JaggedToMD(testingSet), testingOutput);
//Use the most accurate K of KNN
statistics[2, i] = BestKNN;
statistics[3, i] = ParallelRunNN(trainingSet, doubleTO, testingSet, testingOutput);
//RunAnotherNN(trainingSet, doubleTO, testingSet, testingOutput);
}
//Update the classifier lines in the charts
//with the most accurate of the 5 iterations
ChartUpdate(perceChart.Name, "classifier", MostAccuratePerceptron.Item1);
ChartUpdate(nnChart.Name, "classifier1", MostAccurateNN.Item1);
//Process the array with the Confusion Matrices
//and update the list view
var processed = UtilityProvider.ProcessStatistics(statistics);
UpdateStatisticsListView(processed);
}
public void TransformTest()
{
var inputs = yinyang.Submatrix(null, 0, 1).ToArray();
var labels = yinyang.GetColumn(2).ToInt32();
ConfusionMatrix actual, expected;
SequentialMinimalOptimization a, b;
var kernel = new Polynomial(2, 0);
{
var machine = new KernelSupportVectorMachine(kernel, inputs[0].Length);
a = new SequentialMinimalOptimization(machine, inputs, labels);
a.UseComplexityHeuristic = true;
a.Run();
int[] values = new int[labels.Length];
for (int i = 0; i < values.Length; i++)
values[i] = Math.Sign(machine.Compute(inputs[i]));
expected = new ConfusionMatrix(values, labels);
}
{
var projection = inputs.Apply(kernel.Transform);
var machine = new SupportVectorMachine(projection[0].Length);
b = new SequentialMinimalOptimization(machine, projection, labels);
b.UseComplexityHeuristic = true;
b.Run();
int[] values = new int[labels.Length];
for (int i = 0; i < values.Length; i++)
values[i] = Math.Sign(machine.Compute(projection[i]));
actual = new ConfusionMatrix(values, labels);
}
Assert.AreEqual(a.Complexity, b.Complexity, 1e-15);
Assert.AreEqual(expected.TrueNegatives, actual.TrueNegatives);
Assert.AreEqual(expected.TruePositives, actual.TruePositives);
Assert.AreEqual(expected.FalseNegatives, actual.FalseNegatives);
Assert.AreEqual(expected.FalsePositives, actual.FalsePositives);
}
public void MatthewsCorrelationCoefficientTest2()
{
ConfusionMatrix matrix = new ConfusionMatrix(100, 100, 200, 600);
Assert.AreEqual(0.21821789023599236, matrix.MatthewsCorrelationCoefficient);
}
public void WeightRatioTest()
{
var dataset = KernelSupportVectorMachineTest.training;
var inputs = dataset.Submatrix(null, 0, 3);
var labels = Tools.Scale(0, 1, -1, 1, dataset.GetColumn(4)).ToInt32();
Gaussian kernel = Gaussian.Estimate(inputs);
{
var machine = new KernelSupportVectorMachine(kernel, inputs[0].Length);
var smo = new SequentialMinimalOptimization(machine, inputs, labels);
smo.Complexity = 1.0;
smo.WeightRatio = 10;
double error = smo.Run();
Assert.AreEqual(1.0, smo.PositiveWeight);
Assert.AreEqual(0.1, smo.NegativeWeight);
Assert.AreEqual(0.7142857142857143, error);
Assert.AreEqual(265.78327637381551, (machine.Kernel as Gaussian).Sigma);
Assert.AreEqual(39, machine.SupportVectors.Length);
int[] actual = new int[labels.Length];
for (int i = 0; i < actual.Length; i++)
actual[i] = Math.Sign(machine.Compute(inputs[i]));
ConfusionMatrix matrix = new ConfusionMatrix(actual, labels);
Assert.AreEqual(12, matrix.TruePositives); // has more importance
Assert.AreEqual(0, matrix.FalseNegatives); // has more importance
Assert.AreEqual(30, matrix.FalsePositives);
Assert.AreEqual(0, matrix.TrueNegatives);
Assert.AreEqual(1.0, matrix.Sensitivity);
Assert.AreEqual(0.0, matrix.Specificity);
Assert.AreEqual(0.44444444444444448, matrix.FScore);
Assert.AreEqual(0.0, matrix.MatthewsCorrelationCoefficient);
}
{
var machine = new KernelSupportVectorMachine(kernel, inputs[0].Length);
var smo = new SequentialMinimalOptimization(machine, inputs, labels);
smo.Complexity = 1.0;
smo.WeightRatio = 0.1;
double error = smo.Run();
Assert.AreEqual(0.1, smo.PositiveWeight);
Assert.AreEqual(1.0, smo.NegativeWeight);
Assert.AreEqual(0.21428571428571427, error);
Assert.AreEqual(265.78327637381551, (machine.Kernel as Gaussian).Sigma);
Assert.AreEqual(18, machine.SupportVectors.Length);
int[] actual = new int[labels.Length];
for (int i = 0; i < actual.Length; i++)
actual[i] = Math.Sign(machine.Compute(inputs[i]));
ConfusionMatrix matrix = new ConfusionMatrix(actual, labels);
Assert.AreEqual(8, matrix.FalseNegatives);
Assert.AreEqual(1, matrix.FalsePositives); // has more importance
Assert.AreEqual(4, matrix.TruePositives);
Assert.AreEqual(29, matrix.TrueNegatives); // has more importance
Assert.AreEqual(0.33333333333333331, matrix.Sensitivity);
Assert.AreEqual(0.96666666666666667, matrix.Specificity);
Assert.AreEqual(0.47058823529411764, matrix.FScore);
Assert.AreEqual(0.41849149947774944, matrix.MatthewsCorrelationCoefficient);
}
}
public void FixedWeightsTest()
{
var dataset = KernelSupportVectorMachineTest.training;
var inputs = dataset.Submatrix(null, 0, 3);
var labels = Tools.Scale(0, 1, -1, 1, dataset.GetColumn(4)).ToInt32();
KernelSupportVectorMachine machine = new KernelSupportVectorMachine(
Gaussian.Estimate(inputs), inputs[0].Length);
var smo = new SequentialMinimalOptimization(machine, inputs, labels);
smo.Complexity = 10;
double error = smo.Run();
Assert.AreEqual(0.19047619047619047, error);
Assert.AreEqual(265.78327637381551, (machine.Kernel as Gaussian).Sigma);
Assert.AreEqual(29, machine.SupportVectors.Length);
double[] expectedWeights =
{
1.65717694716503, 1.20005456611466, -5.70824245415995, 10,
10, -2.38755497916487, 10, -8.15723436363058, 10, -10, 10,
10, -0.188634936781317, -5.4354281009458, -8.48341139483265,
-5.91105702760141, -5.71489190049223, 10, -2.37289205235858,
-3.33031262413522, -1.97545116517677, 10, -10, -9.563186799279,
-3.917941544845, -0.532584110773336, 4.81951847548326, 0.343668292727091,
-4.34159482731336
};
Assert.IsTrue(expectedWeights.IsEqual(machine.Weights, 1e-6));
int[] actual = new int[labels.Length];
for (int i = 0; i < actual.Length; i++)
actual[i] = Math.Sign(machine.Compute(inputs[i]));
ConfusionMatrix matrix = new ConfusionMatrix(actual, labels);
Assert.AreEqual(8, matrix.FalseNegatives);
Assert.AreEqual(0, matrix.FalsePositives);
Assert.AreEqual(4, matrix.TruePositives);
Assert.AreEqual(30, matrix.TrueNegatives);
Assert.AreEqual(1 / 3.0, matrix.Sensitivity);
Assert.AreEqual(1, matrix.Specificity);
Assert.AreEqual(0.5, matrix.FScore);
Assert.AreEqual(0.5129891760425771, matrix.MatthewsCorrelationCoefficient);
}
public void UseClassProportionsTest()
{
var dataset = KernelSupportVectorMachineTest.training;
var inputs = dataset.Submatrix(null, 0, 3);
var labels = Tools.Scale(0, 1, -1, 1, dataset.GetColumn(4)).ToInt32();
Gaussian kernel = Gaussian.Estimate(inputs);
var machine = new KernelSupportVectorMachine(kernel, inputs[0].Length);
var smo = new SequentialMinimalOptimization(machine, inputs, labels);
smo.Complexity = 1.0;
smo.UseClassProportions = true;
double error = smo.Run();
Assert.AreEqual(1, smo.Complexity);
Assert.AreEqual(0.4, smo.PositiveWeight);
Assert.AreEqual(1.0, smo.NegativeWeight);
Assert.AreEqual(0.4, smo.WeightRatio, 1e-10);
Assert.AreEqual(0.2857142857142857, error);
Assert.AreEqual(265.78327637381551, (machine.Kernel as Gaussian).Sigma);
Assert.AreEqual(26, machine.SupportVectors.Length);
int[] actual = new int[labels.Length];
for (int i = 0; i < actual.Length; i++)
actual[i] = Math.Sign(machine.Compute(inputs[i]));
ConfusionMatrix matrix = new ConfusionMatrix(actual, labels);
Assert.AreEqual(12, matrix.FalseNegatives);
Assert.AreEqual(0, matrix.FalsePositives);
Assert.AreEqual(0, matrix.TruePositives);
Assert.AreEqual(30, matrix.TrueNegatives);
}
public void ToGeneralMatrixTest1()
{
// Example from http://www.iph.ufrgs.br/corpodocente/marques/cd/rd/presabs.htm
ConfusionMatrix matrix = new ConfusionMatrix
(
truePositives: 70,
trueNegatives: 95,
falsePositives: 5,
falseNegatives: 30
);
Assert.AreEqual(70, matrix.TruePositives);
Assert.AreEqual(5, matrix.FalsePositives);
Assert.AreEqual(95, matrix.TrueNegatives);
Assert.AreEqual(30, matrix.FalseNegatives);
GeneralConfusionMatrix general = matrix.ToGeneralMatrix();
Assert.AreEqual(matrix.Kappa, general.Kappa, 1e-10);
Assert.AreEqual(matrix.Accuracy, general.OverallAgreement, 1e-10);
Assert.AreEqual(matrix.StandardError, general.StandardError, 1e-10);
Assert.AreEqual(matrix.StandardErrorUnderNull, general.StandardErrorUnderNull, 1e-10);
Assert.AreEqual(matrix.Variance, general.Variance, 1e-10);
Assert.AreEqual(matrix.VarianceUnderNull, general.VarianceUnderNull, 1e-10);
Assert.AreEqual(matrix.Samples, general.Samples);
Assert.IsFalse(double.IsNaN(general.Kappa));
Assert.IsFalse(double.IsNaN(general.OverallAgreement));
Assert.IsFalse(double.IsNaN(general.StandardError));
Assert.IsFalse(double.IsNaN(general.StandardErrorUnderNull));
Assert.IsFalse(double.IsNaN(general.Variance));
Assert.IsFalse(double.IsNaN(general.VarianceUnderNull));
}
public void ConfusionMatrixConstructorTest3()
{
// System output
int[] predicted = new int[] { 2, 0, 1 };
// Correct output
int[] expected = new int[] { 5, 2, 1 };
// 1 means positive (the others shall be treated as negatives)
int positiveValue = 1;
ConfusionMatrix target = new ConfusionMatrix(predicted, expected, positiveValue);
int falseNegatives = 0;
int falsePositives = 0;
int truePositives = 1;
int trueNegatives = 2;
Assert.AreEqual(predicted.Length, target.Observations);
Assert.AreEqual(2, target.ActualNegatives);
Assert.AreEqual(1, target.ActualPositives);
Assert.AreEqual(2, target.PredictedNegatives);
Assert.AreEqual(1, target.PredictedPositives);
Assert.AreEqual(falseNegatives, target.FalseNegatives);
Assert.AreEqual(falsePositives, target.FalsePositives);
Assert.AreEqual(truePositives, target.TruePositives);
Assert.AreEqual(trueNegatives, target.TrueNegatives);
Assert.AreEqual(1.0, target.Accuracy);
Assert.AreEqual(1.0, target.Sensitivity);
Assert.AreEqual(1.0, target.Specificity);
Assert.AreEqual(1.0, target.Efficiency);
// Perfect prediction
Assert.AreEqual(1.0, target.MatthewsCorrelationCoefficient);
}
public void ComputeTest5()
{
var dataset = SequentialMinimalOptimizationTest.yinyang;
double[][] inputs = dataset.Submatrix(null, 0, 1).ToArray();
int[] labels = dataset.GetColumn(2).ToInt32();
var kernel = new Polynomial(2, 1);
Accord.Math.Tools.SetupGenerator(0);
var projection = inputs.Apply(kernel.Transform);
var machine = new SupportVectorMachine(projection[0].Length);
var smo = new LinearCoordinateDescent(machine, projection, labels)
{
Complexity = 1000000,
Tolerance = 1e-15
};
double error = smo.Run();
Assert.AreEqual(1000000.0, smo.Complexity, 1e-15);
int[] actual = new int[labels.Length];
for (int i = 0; i < actual.Length; i++)
actual[i] = Math.Sign(machine.Compute(projection[i]));
ConfusionMatrix matrix = new ConfusionMatrix(actual, labels);
Assert.AreEqual(6, matrix.FalseNegatives);
Assert.AreEqual(7, matrix.FalsePositives);
Assert.AreEqual(44, matrix.TruePositives);
Assert.AreEqual(43, matrix.TrueNegatives);
}
private void btnTestingRun_Click(object sender, EventArgs e)
{
if (ann == null || dgvTestingSource.DataSource == null)
{
MessageBox.Show("Please create a machine first.");
return;
}
// Creates a matrix from the source data table
double[,] sourceMatrix = (dgvTestingSource.DataSource as DataTable).ToMatrix();
// Extract inputs
double[][] inputs = new double[sourceMatrix.GetLength(0)][];
for (int i = 0; i < inputs.Length; i++)
inputs[i] = new double[] { sourceMatrix[i, 0], sourceMatrix[i, 1] };
// Get only the label outputs
int[] expected = new int[sourceMatrix.GetLength(0)];
for (int i = 0; i < expected.Length; i++)
expected[i] = (int)sourceMatrix[i, 2];
// Compute the machine outputs
int[] output = new int[expected.Length];
for (int i = 0; i < expected.Length; i++)
output[i] = System.Math.Sign(ann.Compute(inputs[i])[0]);
double[] expectedd = new double[expected.Length];
double[] outputd = new double[expected.Length];
for (int i = 0; i < expected.Length; i++)
{
expectedd[i] = expected[i];
outputd[i] = output[i];
}
// Use confusion matrix to compute some statistics.
ConfusionMatrix confusionMatrix = new ConfusionMatrix(output, expected, 1, -1);
dgvPerformance.DataSource = new List<ConfusionMatrix> { confusionMatrix };
foreach (DataGridViewColumn col in dgvPerformance.Columns) col.Visible = true;
Column1.Visible = Column2.Visible = false;
// Create performance scatterplot
CreateResultScatterplot(zedGraphControl1, inputs, expectedd, outputd);
}
public void ComputeTest5()
{
var dataset = yinyang;
double[][] inputs = dataset.Submatrix(null, 0, 1).ToArray();
int[] labels = dataset.GetColumn(2).ToInt32();
{
Linear kernel = new Linear();
var machine = new KernelSupportVectorMachine(kernel, inputs[0].Length);
var smo = new SequentialMinimalOptimization(machine, inputs, labels);
smo.Complexity = 1.0;
double error = smo.Run();
Assert.AreEqual(1.0, smo.Complexity);
Assert.AreEqual(1.0, smo.WeightRatio);
Assert.AreEqual(1.0, smo.NegativeWeight);
Assert.AreEqual(1.0, smo.PositiveWeight);
Assert.AreEqual(0.14, error);
Assert.AreEqual(30, machine.SupportVectors.Length);
double[] actualWeights = machine.Weights;
double[] expectedWeights = { -1, -1, 1, -1, -1, 1, 1, -1, 1, -1, 1, 1, -1, 0.337065120144639, -1, 1, -0.337065120144639, -1, 1, 1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1 };
Assert.IsTrue(expectedWeights.IsEqual(actualWeights, 1e-10));
int[] actual = new int[labels.Length];
for (int i = 0; i < actual.Length; i++)
actual[i] = Math.Sign(machine.Compute(inputs[i]));
ConfusionMatrix matrix = new ConfusionMatrix(actual, labels);
Assert.AreEqual(7, matrix.FalseNegatives);
Assert.AreEqual(7, matrix.FalsePositives);
Assert.AreEqual(43, matrix.TruePositives);
Assert.AreEqual(43, matrix.TrueNegatives);
}
{
Linear kernel = new Linear();
var machine = new KernelSupportVectorMachine(kernel, inputs[0].Length);
var smo = new SequentialMinimalOptimization(machine, inputs, labels);
smo.Complexity = 1.0;
smo.PositiveWeight = 0.3;
smo.NegativeWeight = 1.0;
double error = smo.Run();
Assert.AreEqual(1.0, smo.Complexity);
Assert.AreEqual(0.3 / 1.0, smo.WeightRatio);
Assert.AreEqual(1.0, smo.NegativeWeight);
Assert.AreEqual(0.3, smo.PositiveWeight);
Assert.AreEqual(0.21, error);
Assert.AreEqual(24, machine.SupportVectors.Length);
double[] actualWeights = machine.Weights;
//string str = actualWeights.ToString(Accord.Math.Formats.CSharpArrayFormatProvider.InvariantCulture);
double[] expectedWeights = { -0.771026323762095, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -0.928973676237905, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1 };
Assert.IsTrue(expectedWeights.IsEqual(actualWeights, 1e-10));
int[] actual = new int[labels.Length];
for (int i = 0; i < actual.Length; i++)
actual[i] = (int)machine.Compute(inputs[i]);
ConfusionMatrix matrix = new ConfusionMatrix(actual, labels);
Assert.AreEqual(50, matrix.FalseNegatives);
Assert.AreEqual(0, matrix.FalsePositives);
Assert.AreEqual(0, matrix.TruePositives);
Assert.AreEqual(50, matrix.TrueNegatives);
}
{
Linear kernel = new Linear();
var machine = new KernelSupportVectorMachine(kernel, inputs[0].Length);
var smo = new SequentialMinimalOptimization(machine, inputs, labels);
smo.Complexity = 1.0;
smo.PositiveWeight = 1.0;
smo.NegativeWeight = 0.3;
double error = smo.Run();
Assert.AreEqual(1.0, smo.Complexity);
Assert.AreEqual(1.0 / 0.3, smo.WeightRatio);
Assert.AreEqual(0.3, smo.NegativeWeight);
Assert.AreEqual(1.0, smo.PositiveWeight);
Assert.AreEqual(0.15, error);
Assert.AreEqual(19, machine.SupportVectors.Length);
double[] actualWeights = machine.Weights;
double[] expectedWeights = new double[] { 1, 1, -0.3, 1, -0.3, 1, 1, -0.3, 1, 1, 1, 1, 1, 1, 1, 1, 0.129080057278249, 1, 0.737797469918795 };
Assert.IsTrue(expectedWeights.IsEqual(actualWeights, 1e-10));
int[] actual = new int[labels.Length];
for (int i = 0; i < actual.Length; i++)
actual[i] = Math.Sign(machine.Compute(inputs[i]));
ConfusionMatrix matrix = new ConfusionMatrix(actual, labels);
Assert.AreEqual(0, matrix.FalseNegatives);
Assert.AreEqual(50, matrix.FalsePositives);
Assert.AreEqual(50, matrix.TruePositives);
Assert.AreEqual(0, matrix.TrueNegatives);
}
}
public void ComputeTest5()
{
var dataset = SequentialMinimalOptimizationTest.yinyang;
var inputs = dataset.Submatrix(null, 0, 1).ToArray();
var labels = dataset.GetColumn(2).ToInt32();
var kernel = new Polynomial(2, 0);
{
var machine = new KernelSupportVectorMachine(kernel, inputs[0].Length);
var smo = new SequentialMinimalOptimization(machine, inputs, labels);
smo.UseComplexityHeuristic = true;
double error = smo.Run();
Assert.AreEqual(0.2, error);
Assert.AreEqual(0.11714451552090824, smo.Complexity);
int[] actual = new int[labels.Length];
for (int i = 0; i < actual.Length; i++)
actual[i] = Math.Sign(machine.Compute(inputs[i]));
ConfusionMatrix matrix = new ConfusionMatrix(actual, labels);
Assert.AreEqual(20, matrix.FalseNegatives);
Assert.AreEqual(0, matrix.FalsePositives);
Assert.AreEqual(30, matrix.TruePositives);
Assert.AreEqual(50, matrix.TrueNegatives);
}
{
Accord.Math.Tools.SetupGenerator(0);
var projection = inputs.Apply(kernel.Transform);
var machine = new SupportVectorMachine(projection[0].Length);
var smo = new LinearNewtonMethod(machine, projection, labels);
smo.UseComplexityHeuristic = true;
double error = smo.Run();
Assert.AreEqual(0.18, error);
Assert.AreEqual(0.11714451552090821, smo.Complexity, 1e-15);
int[] actual = new int[labels.Length];
for (int i = 0; i < actual.Length; i++)
actual[i] = Math.Sign(machine.Compute(projection[i]));
ConfusionMatrix matrix = new ConfusionMatrix(actual, labels);
Assert.AreEqual(17, matrix.FalseNegatives);
Assert.AreEqual(1, matrix.FalsePositives);
Assert.AreEqual(33, matrix.TruePositives);
Assert.AreEqual(49, matrix.TrueNegatives);
}
}
public void Start()
{
//Read the data from the files
var file1DataRaw = AlgorithmHelpers.ReadMatrixFromFile(@"class_1.dat");
var file2DataRaw = AlgorithmHelpers.ReadMatrixFromFile(@"class_2.dat");
ClassA = AlgorithmHelpers.ScaleDown(AlgorithmHelpers.ChooseFeatures(file1DataRaw));
ClassB = AlgorithmHelpers.ScaleDown(AlgorithmHelpers.ChooseFeatures(file2DataRaw));
m_maxX = ClassA.Max(0)[0] > ClassB.Max(0)[0] ? ClassA.Max(0)[0] : ClassB.Max(0)[0];
m_minX = ClassA.Min(0)[0] < ClassB.Min(0)[0] ? ClassA.Min(0)[0] : ClassB.Min(0)[0];
//Fill the charts with the data
m_view.ShowTrainingData(ClassA, ClassB);
//Clear the list view
m_view.ClearListView();
//Fill it with the confusion matrix for each algorithm per iteration
var statistics = new ConfusionMatrix[4, 5];
//Merge the two data sets
//and run kmeans
var kmeans = new KMeansClustering(AlgorithmHelpers.MergeArrays(file1DataRaw, file2DataRaw),
m_view.Iterations,
m_view.ThetaStep);
var idx = kmeans.Classify();
m_view.ClustersTextUpdate(idx.Distinct().Length.ToString());
m_view.ZeroProgressBar();
m_view.StepProgressBar();
//Partition m_iterations times and run the algorithms
for (int i = 0; i < m_iterations; ++i)
{
m_view.PerformStep();
//Partition its class to training and testing set
var partitions = new DataPartition[] { AlgorithmHelpers.Partition(ClassA), AlgorithmHelpers.Partition(ClassB) };
//Create the training data
var trainingPair = AlgorithmHelpers.CreateDataPair(partitions[0].Item1, partitions[1].Item1);
var trainingSet = trainingPair.Item1;
var trainingOutput = trainingPair.Item2;
//Create the testing data
var testingPair = AlgorithmHelpers.CreateDataPair(partitions[0].Item2, partitions[1].Item2);
var testingSet = testingPair.Item1;
var testingOutput = testingPair.Item2;
//Some functions need the training output to be a vector of doubles
var doubleTO = trainingOutput
.Select(x => new[] { Convert.ToDouble(x) })
.ToArray();
for (int k = 1; k < 3; ++k)
{
var nn = new KNearestNeighboursRuntime(k, trainingSet, trainingOutput, testingSet, testingOutput);
if (BestKNN == null)
{
BestKNN = nn.Execute();
}
else
{
var iter = nn.Execute();
if (iter.Accuracy > BestKNN.Accuracy)
BestKNN = iter;
}
}
var perceptron = new PerceptronRuntime(trainingSet, doubleTO, testingSet, testingOutput);
perceptron.Finished += perceptron_Finished;
var leastSquare = new LeastSquaresRuntime(AlgorithmHelpers.JaggedToMD(trainingSet), AlgorithmHelpers.JaggedToMD(doubleTO), AlgorithmHelpers.JaggedToMD(testingSet), testingOutput);
var neuralNetwork = new ParallelNeuralNetworkRuntime(trainingSet, doubleTO, testingSet, testingOutput);
neuralNetwork.Finished += neuralNetwork_Finished;
//Compute the confusion matrices for the four classifiers
statistics[0, i] = perceptron.Execute();
statistics[1, i] = leastSquare.Execute();
//Use the most accurate K of KNN
statistics[2, i] = BestKNN;
statistics[3, i] = neuralNetwork.Execute();
}
//Update the classifier lines in the charts
//with the most accurate of the 5 iterations
m_view.ChartUpdate("", "classifier", MostAccuratePerceptron.Item1);
m_view.ChartUpdate("", "classifier1", MostAccurateNN.Item1);
//Process the array with the Confusion Matrices
//and update the list view
var processed = AlgorithmHelpers.ProcessStatistics(statistics);
m_view.UpdateStatisticsListView(processed);
}
/// <summary>
/// Tests the previously created tree into a new set of data.
/// </summary>
///
private void btnTestingRun_Click(object sender, EventArgs e)
{
if (tree == null || dgvTestingSource.DataSource == null)
{
MessageBox.Show("Please create a machine first.");
return;
}
// Creates a matrix from the entire source data table
double[,] table = (dgvLearningSource.DataSource as DataTable).ToMatrix(out columnNames);
// Get only the input vector values (first two columns)
double[][] inputs = table.GetColumns(0, 1).ToArray();
// Get the expected output labels (last column)
int[] expected = table.GetColumn(2).ToInt32();
// Compute the actual tree outputs
int[] actual = new int[inputs.Length];
for (int i = 0; i < inputs.Length; i++)
actual[i] = tree.Compute(inputs[i]);
// Use confusion matrix to compute some statistics.
ConfusionMatrix confusionMatrix = new ConfusionMatrix(actual, expected, 1, 0);
dgvPerformance.DataSource = new [] { confusionMatrix };
// Create performance scatter plot
CreateResultScatterplot(zedGraphControl1, inputs, expected.ToDouble(), actual.ToDouble());
}
public void ConfusionMatrixConstructorTest2()
{
// The correct and expected output values (as confirmed by a Gold
// standard rule, actual experiment or true verification)
bool[] expected = { false, false, true, false, true, false, false, false, false, false };
// The values as predicted by the decision system or
// the test whose performance is being measured.
bool[] predicted = { false, false, false, true, true, false, false, false, false, true };
// Create a new confusion matrix using the given parameters
ConfusionMatrix matrix = new ConfusionMatrix(predicted, expected);
int falseNegatives = 1;
int falsePositives = 2;
int truePositives = 1;
int trueNegatives = 6;
Assert.AreEqual(predicted.Length, matrix.Observations);
Assert.AreEqual(8, matrix.ActualNegatives);
Assert.AreEqual(2, matrix.ActualPositives);
Assert.AreEqual(7, matrix.PredictedNegatives);
Assert.AreEqual(3, matrix.PredictedPositives);
Assert.AreEqual(falseNegatives, matrix.FalseNegatives);
Assert.AreEqual(falsePositives, matrix.FalsePositives);
Assert.AreEqual(truePositives, matrix.TruePositives);
Assert.AreEqual(trueNegatives, matrix.TrueNegatives);
}
public void WeightsTest1()
{
var dataset = yinyang;
double[][] inputs = dataset.Submatrix(null, 0, 1).ToArray();
int[] labels = dataset.GetColumn(2).ToInt32();
Accord.Math.Tools.SetupGenerator(0);
var kernel = new Linear(1);
{
var machine = new KernelSupportVectorMachine(kernel, inputs[0].Length);
var smo = new SequentialMinimalOptimization(machine, inputs, labels);
smo.Complexity = 1.0;
smo.PositiveWeight = 1;
smo.NegativeWeight = 1;
smo.Tolerance = 0.001;
double error = smo.Run();
int[] actual = new int[labels.Length];
for (int i = 0; i < actual.Length; i++)
actual[i] = Math.Sign(machine.Compute(inputs[i]));
ConfusionMatrix matrix = new ConfusionMatrix(actual, labels);
Assert.AreEqual(43, matrix.TruePositives); // both classes are
Assert.AreEqual(43, matrix.TrueNegatives); // well equilibrated
Assert.AreEqual(7, matrix.FalseNegatives);
Assert.AreEqual(7, matrix.FalsePositives);
Assert.AreEqual(1.0, smo.Complexity);
Assert.AreEqual(1.0, smo.WeightRatio);
Assert.AreEqual(1.0, smo.NegativeWeight);
Assert.AreEqual(1.0, smo.PositiveWeight);
Assert.AreEqual(0.14, error);
Assert.AreEqual(0.001, smo.Tolerance);
Assert.AreEqual(31, machine.SupportVectors.Length);
}
{
var machine = new KernelSupportVectorMachine(kernel, inputs[0].Length);
var smo = new SequentialMinimalOptimization(machine, inputs, labels);
smo.Complexity = 1;
smo.PositiveWeight = 100;
smo.NegativeWeight = 1;
smo.Tolerance = 0.001;
double error = smo.Run();
int[] actual = new int[labels.Length];
for (int i = 0; i < actual.Length; i++)
actual[i] = Math.Sign(machine.Compute(inputs[i]));
ConfusionMatrix matrix = new ConfusionMatrix(actual, labels);
Assert.AreEqual(50, matrix.TruePositives); // has more importance
Assert.AreEqual(23, matrix.TrueNegatives);
Assert.AreEqual(0, matrix.FalseNegatives); // has more importance
Assert.AreEqual(27, matrix.FalsePositives);
Assert.AreEqual(1.0, smo.Complexity);
Assert.AreEqual(100, smo.WeightRatio);
Assert.AreEqual(1.0, smo.NegativeWeight);
Assert.AreEqual(100, smo.PositiveWeight);
Assert.AreEqual(0.001, smo.Tolerance);
Assert.AreEqual(0.27, error);
Assert.AreEqual(41, machine.SupportVectors.Length);
}
{
var machine = new KernelSupportVectorMachine(kernel, inputs[0].Length);
var smo = new SequentialMinimalOptimization(machine, inputs, labels);
smo.Complexity = 1;
smo.PositiveWeight = 1;
smo.NegativeWeight = 100;
smo.Tolerance = 0.001;
double error = smo.Run();
int[] actual = new int[labels.Length];
for (int i = 0; i < actual.Length; i++)
actual[i] = Math.Sign(machine.Compute(inputs[i]));
var matrix = new ConfusionMatrix(actual, labels);
Assert.AreEqual(25, matrix.TruePositives);
Assert.AreEqual(50, matrix.TrueNegatives); // has more importance
Assert.AreEqual(25, matrix.FalseNegatives);
Assert.AreEqual(0, matrix.FalsePositives); // has more importance
Assert.AreEqual(1.0, smo.Complexity);
Assert.AreEqual(0.01, smo.WeightRatio);
Assert.AreEqual(100, smo.NegativeWeight);
Assert.AreEqual(1.0, smo.PositiveWeight);
Assert.AreEqual(0.25, error);
Assert.AreEqual(0.001, smo.Tolerance);
Assert.AreEqual(40, machine.SupportVectors.Length);
}
}
public void ConfusionMatrixConstructorTest4()
{
// Example from http://www.iph.ufrgs.br/corpodocente/marques/cd/rd/presabs.htm
ConfusionMatrix matrix = new ConfusionMatrix(70, 95, 5, 30);
// Prevalence 0.500 0.100 0.011
// Overall Power 0.500 0.900 0.989
// Sensitivity 0.700 0.700 0.700
// Specificity 0.950 0.950 0.950
// PPP 0.933 0.610 0.130
// NPP 0.760 0.970 0.997
// Misc. Rate 0.175 0.075 0.053
// Odds Ratio 44.333 44.333 44.333
// Kappa 0.650 0.610 0.210
// NMI 0.371 0.360 0.264
Assert.AreEqual(0.500, matrix.OverallDiagnosticPower, 1e-3);
Assert.AreEqual(0.700, matrix.Sensitivity, 1e-3);
Assert.AreEqual(0.950, matrix.Specificity, 1e-3);
Assert.AreEqual(0.933, matrix.PositivePredictiveValue, 1e-3);
Assert.AreEqual(0.760, matrix.NegativePredictiveValue, 1e-3);
Assert.AreEqual(0.175, 1 - matrix.Accuracy, 1e-3);
Assert.AreEqual(44.333, matrix.OddsRatio, 1e-3);
Assert.AreEqual(0.650, matrix.Kappa, 1e-3);
Assert.AreEqual(0.371, matrix.NormalizedMutualInformation, 1e-3);
Assert.IsFalse(double.IsNaN(matrix.OverallDiagnosticPower));
Assert.IsFalse(double.IsNaN(matrix.Sensitivity));
Assert.IsFalse(double.IsNaN(matrix.Specificity));
Assert.IsFalse(double.IsNaN(matrix.PositivePredictiveValue));
Assert.IsFalse(double.IsNaN(matrix.NegativePredictiveValue));
Assert.IsFalse(double.IsNaN(matrix.Accuracy));
Assert.IsFalse(double.IsNaN(matrix.OddsRatio));
Assert.IsFalse(double.IsNaN(matrix.Kappa));
Assert.IsFalse(double.IsNaN(matrix.NormalizedMutualInformation));
}
private static void testWeights(double[][] inputs, int[] labels, IKernel kernel)
{
{
var machine = new KernelSupportVectorMachine(kernel, inputs[0].Length);
var smo = new SequentialMinimalOptimization(machine, inputs, labels);
smo.PositiveWeight = 100;
smo.NegativeWeight = 1;
double error = smo.Run();
int[] actual = new int[labels.Length];
for (int i = 0; i < actual.Length; i++)
actual[i] = Math.Sign(machine.Compute(inputs[i]));
ConfusionMatrix matrix = new ConfusionMatrix(actual, labels);
Assert.AreEqual(50, matrix.TruePositives); // has more importance
Assert.AreEqual(0, matrix.FalseNegatives); // has more importance
}
{
var machine = new KernelSupportVectorMachine(kernel, inputs[0].Length);
var smo = new SequentialMinimalOptimization(machine, inputs, labels);
smo.PositiveWeight = 1;
smo.NegativeWeight = 100;
double error = smo.Run();
int[] actual = new int[labels.Length];
for (int i = 0; i < actual.Length; i++)
actual[i] = Math.Sign(machine.Compute(inputs[i]));
var matrix = new ConfusionMatrix(actual, labels);
Assert.AreEqual(50, matrix.TrueNegatives); // has more importance
Assert.AreEqual(0, matrix.FalsePositives); // has more importance
}
}
