C# (CSharp) MulticlassLibSVM Exemples

Exemple #1

Fichier : classifier_multiclasslibsvm_modular.cs Projet : joseph-chan/rqpersonalsvn

    public static void Main()
    {
        modshogun.init_shogun_with_defaults();
        double width = 2.1;
        double epsilon = 1e-5;
        double C = 1.0;

        double[,] traindata_real = Load.load_numbers("../data/fm_train_real.dat");
        double[,] testdata_real = Load.load_numbers("../data/fm_test_real.dat");

        double[] trainlab = Load.load_labels("../data/label_train_multiclass.dat");

        RealFeatures feats_train = new RealFeatures();
        feats_train.set_feature_matrix(traindata_real);
        RealFeatures feats_test = new RealFeatures();
        feats_test.set_feature_matrix(testdata_real);

        GaussianKernel kernel = new GaussianKernel(feats_train, feats_train, width);

        MulticlassLabels labels = new MulticlassLabels(trainlab);

        MulticlassLibSVM svm = new MulticlassLibSVM(C, kernel, labels);
        svm.set_epsilon(epsilon);
        svm.train();

        kernel.init(feats_train, feats_test);
        double[] out_labels = MulticlassLabels.obtain_from_generic(svm.apply()).get_labels();

        foreach (double item in out_labels)
            Console.Write(item);

        modshogun.exit_shogun();
    }

Exemple #2

Fichier : classifier_multiclasslibsvm_modular.cs Projet : serialhex/shogun

    public static void Main()
    {
        modshogun.init_shogun_with_defaults();
        double width   = 2.1;
        double epsilon = 1e-5;
        double C       = 1.0;

        double[,] traindata_real = Load.load_numbers("../data/fm_train_real.dat");
        double[,] testdata_real  = Load.load_numbers("../data/fm_test_real.dat");

        double[] trainlab = Load.load_labels("../data/label_train_multiclass.dat");

        RealFeatures feats_train = new RealFeatures();

        feats_train.set_feature_matrix(traindata_real);
        RealFeatures feats_test = new RealFeatures();

        feats_test.set_feature_matrix(testdata_real);

        GaussianKernel kernel = new GaussianKernel(feats_train, feats_train, width);

        Labels labels = new Labels(trainlab);

        MulticlassLibSVM svm = new MulticlassLibSVM(C, kernel, labels);

        svm.set_epsilon(epsilon);
        svm.train();

        kernel.init(feats_train, feats_test);
        double[] out_labels = svm.apply().get_labels();

        foreach (double item in out_labels)
        {
            Console.Write(item);
        }

        modshogun.exit_shogun();
    }