def validate(name, X, Y, normal_class): anomaly = [x for x, y in zip(X, Y) if y != normal_class] y_anomaly = [y for y in Y if y != normal_class] normal = [x for x, y in zip(X, Y) if y == normal_class] y_normal = [y for y in Y if y == normal_class] nu_range = np.logspace(-3, 0, 30) * 0.2 gamma_range = np.logspace(-5, 4, 10) validator = Validator(X, Y, normal, y_normal, anomaly, y_anomaly) """ hypersphere_valid = validator.hyperspherical_predictor(name, nu_range, gamma_range) print hypersphere_valid[:-3] one_class_svm_valid = validator.one_class_svm(name, nu_range, gamma_range) print one_class_svm_valid[:-3] plt.plot(hypersphere_valid[-3], hypersphere_valid[-2], 'b') plt.plot(hypersphere_valid[-3], hypersphere_valid[-1], 'b--') plt.plot(one_class_svm_valid[-3], one_class_svm_valid[-2], 'g') plt.plot(one_class_svm_valid[-3], one_class_svm_valid[-1], 'g--') #plt.show() plt.savefig('graphs/' + name + '.png') """ print validator.svc_biclass(name, nu_range, gamma_range) print validator.multiclass_hyperspherical_predictor(name, nu_range, gamma_range) print validator.multiclass_one_class_svm(name, nu_range, gamma_range) print validator.svc(name, nu_range, gamma_range)