public static SupportVectorMachine <Gaussian> MakeDeserialization(string path)
{
var teacher = new SequentialMinimalOptimization <Gaussian>()
{
UseComplexityHeuristic = true,
UseKernelEstimation = true
};
double[][] inputs2 = new double[4][];
inputs2[0] = new[] { 1.0, 4 };
inputs2[1] = new[] { 6.0, 8 };
inputs2[2] = new[] { 60.0, 78 };
inputs2[3] = new[] { 60.0, 90 };
double[] outputs2 = { 1, 1, 0, 0 };
SupportVectorMachine <Gaussian> svmAfter = teacher.Learn(inputs2, outputs2);
SVMGaussianData dataAfter = Deserialize <SVMGaussianData>(path);
svmAfter.NumberOfInputs = dataAfter.NumberOfInputs;
svmAfter.NumberOfOutputs = dataAfter.NumberOfOutputs;
svmAfter.SupportVectors = dataAfter.SupportVectors;
svmAfter.Threshold = dataAfter.Threshold;
svmAfter.Weights = dataAfter.Weights;
Gaussian g = new Gaussian();
g.Gamma = dataAfter.Gamma;
g.Sigma = dataAfter.Sigma;
g.SigmaSquared = dataAfter.SigmaSquared;
svmAfter.Kernel = g;
return(svmAfter);
}
public static void MakeSerialization(SupportVectorMachine <Gaussian> svm, string path)
{
SVMGaussianData data = new SVMGaussianData();
data.NumberOfInputs = svm.NumberOfInputs;
data.NumberOfOutputs = svm.NumberOfOutputs;
data.SupportVectors = svm.SupportVectors;
data.Threshold = svm.Threshold;
data.Weights = svm.Weights;
data.Sigma = svm.Kernel.Sigma;
data.SigmaSquared = svm.Kernel.SigmaSquared;
data.Gamma = svm.Kernel.Gamma;
Serialize(data, path);
}
