def __init__(self, source_dataset, target_dataset, H=None, alpha=0.82, maxiter=30, tol=1e-4):
self.source_dataset = source_dataset
self.target_dataset = target_dataset
self.alignment_matrix = None
self.A1 = source_dataset.get_adjacency_matrix()
self.A2 = target_dataset.get_adjacency_matrix()
self.alpha = alpha
self.maxiter = maxiter
self.H = get_H(H, source_dataset, target_dataset)
self.tol = tol
self.N1 = source_dataset.features
self.N2 = target_dataset.features
self.E1 = source_dataset.load_edge_features()
self.E2 = target_dataset.load_edge_features()
self.alignment_matrix = None
# % If no node attributes input, then initialize as a vector of 1
# % so that all nodes are treated to have the save attributes which
# % is equivalent to no given node attribute.
# E1 should be (2, A1.shape[0], A1.shape[1])
if self.N1 is None and self.N2 is None:
self.N1 = np.ones((self.A1.shape[0], 1))
self.N2 = np.ones((self.A2.shape[0], 1))
if self.E1 is None and self.E2 is None:
self.E1 = np.zeros((1, self.A1.shape[0], self.A1.shape[1]))
self.E2 = np.zeros((1, self.A2.shape[0], self.A2.shape[1]))
self.E1[0] = self.A1
self.E2[0] = self.A2
def __init__(self,
source_dataset,
target_dataset,
H=None,
alpha=0.82,
maxiter=30,
tol=1e-4):
self.source_dataset = source_dataset
self.target_dataset = target_dataset
self.alignment_matrix = None
self.A1 = source_dataset.get_adjacency_matrix()
self.A2 = target_dataset.get_adjacency_matrix()
self.alpha = alpha
self.maxiter = maxiter
self.H = get_H(H, source_dataset, target_dataset)
self.tol = tol
def __init__(self,
source_dataset,
target_dataset,
H=None,
alpha=0.82,
maxiter=30,
tol=1e-4,
train_dict=None):
self.source_dataset = source_dataset
self.target_dataset = target_dataset
self.alignment_matrix = None
self.A1 = source_dataset.get_adjacency_matrix()
self.A2 = target_dataset.get_adjacency_matrix()
self.alpha = alpha
self.maxiter = maxiter
if train_dict is not None:
print("This is supervised IsoRank")
self.H = get_H(H, source_dataset, target_dataset, train_dict)
self.tol = tol