A_sexage = preprocessing_dataset.str_to_float(A3) labels, _, _ = read_tadpole.load_csv_no_header(labels_path) labels = preprocessing_dataset.str_to_float(labels) M_float = preprocessing_dataset.preprocessing_nan_normalization(M_str) imp = Imputer(missing_values='NaN', strategy='mean', axis=0) imp = imp.fit(M_float) # Impute our data, then train M_float_imp = imp.transform(M_float) M = M_float_imp A_age = preprocessing_dataset.normalize_adj(A_age) A_sex = preprocessing_dataset.normalize_adj(A_sex) A_sexage = preprocessing_dataset.normalize_adj(A_sexage) mask_age = preprocessing_dataset.str_to_float(mask_age) mask_sex = preprocessing_dataset.str_to_float(mask_sex) mask_agesex = preprocessing_dataset.str_to_float(mask_agesex) mask_nosignificance = preprocessing_dataset.str_to_float(mask_nosignificance) #computation of the normalized laplacians Lrow_age = csgraph.laplacian(A_age, normed=True) Lrow_sex = csgraph.laplacian(A_sex, normed=True) Lrow_agesex = csgraph.laplacian(A_sexage, normed=True) class Train_test_matrix_completion:
Wrow, _, _ = read_tadpole.load_csv_no_header(path_dataset_affinity_matrix)
labels, _, _ = read_tadpole.load_csv_no_header(labels_path)
#parameters/preprocessing step that do not change during the running
Wrow = preprocessing_dataset.str_to_float(Wrow)
labels = preprocessing_dataset.str_to_float(labels)
M_float = preprocessing_dataset.preprocessing_nan_normalization(M_str)
imp = Imputer(missing_values='NaN', strategy='mean', axis=0)
imp = imp.fit(M_float)
# Impute our data
M_float_imp = imp.transform(M_float)
M = M_float_imp
Wrow = preprocessing_dataset.normalize_adj(Wrow)
#computation of the normalized laplacians
Lrow = csgraph.laplacian(Wrow, normed=True)
ord_row = 3 # row for the Chebyshev polynomials
class Train_test_matrix_completion:
"""
The neural network model.
"""
def frobenius_norm_square(self, tensor):
"""
Function that returns the squared Frobenius norm of tensor
Input: tensor: the tensor that we would like to know the norm of
Output: Frobenius norm of the tensor
