bool CLinearRidgeRegression::train_machine(CFeatures* data)
{
REQUIRE(m_labels,"No labels set\n");
if (!data)
data=features;
REQUIRE(data,"No features provided and no featured previously set\n");
REQUIRE(m_labels->get_num_labels() == data->get_num_vectors(),
"Number of training vectors (%d) does not match number of labels (%d)\n",
m_labels->get_num_labels(), data->get_num_vectors());
REQUIRE(data->get_feature_class() == C_DENSE,
"Expected Dense Features (%d) but got (%d)\n",
C_DENSE, data->get_feature_class());
REQUIRE(data->get_feature_type() == F_DREAL,
"Expected Real Features (%d) but got (%d)\n",
F_DREAL, data->get_feature_type());
CDenseFeatures<float64_t>* feats=(CDenseFeatures<float64_t>*) data;
int32_t num_feat=feats->get_num_features();
int32_t num_vec=feats->get_num_vectors();
SGMatrix<float64_t> kernel_matrix(num_feat,num_feat);
SGMatrix<float64_t> feats_matrix(feats->get_feature_matrix());
SGVector<float64_t> y(num_feat);
SGVector<float64_t> tau_vector(num_feat);
tau_vector.zero();
tau_vector.add(m_tau);
Map<MatrixXd> eigen_kernel_matrix(kernel_matrix.matrix, num_feat,num_feat);
Map<MatrixXd> eigen_feats_matrix(feats_matrix.matrix, num_feat,num_vec);
Map<VectorXd> eigen_y(y.vector, num_feat);
Map<VectorXd> eigen_labels(((CRegressionLabels*)m_labels)->get_labels(),num_vec);
Map<VectorXd> eigen_tau(tau_vector.vector, num_feat);
eigen_kernel_matrix = eigen_feats_matrix*eigen_feats_matrix.transpose();
eigen_kernel_matrix.diagonal() += eigen_tau;
eigen_y = eigen_feats_matrix*eigen_labels ;
LLT<MatrixXd> llt;
llt.compute(eigen_kernel_matrix);
if(llt.info() != Eigen::Success)
{
SG_WARNING("Features covariance matrix was not positive definite\n");
return false;
}
eigen_y = llt.solve(eigen_y);
set_w(y);
return true;
}
bool CKernelRidgeRegression::solve_krr_system()
{
SGMatrix<float64_t> kernel_matrix(kernel->get_kernel_matrix());
int32_t n = kernel_matrix.num_rows;
SGVector<float64_t> y = ((CRegressionLabels*)m_labels)->get_labels();
for(index_t i=0; i<n; i++)
kernel_matrix(i,i) += m_tau;
Map<MatrixXd> eigen_kernel_matrix(kernel_matrix.matrix, n, n);
Map<VectorXd> eigen_alphas(m_alpha.vector, n);
Map<VectorXd> eigen_y(y.vector, n);
LLT<MatrixXd> llt;
llt.compute(eigen_kernel_matrix);
if (llt.info() != Eigen::Success)
{
SG_WARNING("Features covariance matrix was not positive definite\n");
return false;
}
eigen_alphas = llt.solve(eigen_y);
return true;
}