def test_neighborhood_subgraph_pairwise_distance():
"""Eigenvalue test for the neighborhood subgraph pairwise distance kernel."""
nspd_kernel = NeighborhoodSubgraphPairwiseDistance(verbose=verbose,
normalize=normalize)
if verbose:
print_kernel("NSPD", nspd_kernel, dataset_tr, dataset_te)
else:
positive_eig(nspd_kernel, dataset)
def test_neighborhood_subgraph_pairwise_distance():
"""Picklability test for the Neighborhood Subgraph Pairwise Distance kernel [+ generic-wrapper]."""
train, _ = generate_dataset(n_graphs=100,
r_vertices=(5, 10),
r_connectivity=(0.4, 0.8),
r_weight_edges=(1, 1),
n_graphs_test=40,
random_state=rs,
features=('nl', 5, 'el', 4))
nspd_kernel = NeighborhoodSubgraphPairwiseDistance(verbose=verbose,
normalize=normalize)
gk = GraphKernel(
kernel={"name": "neighborhood_subgraph_pairwise_distance"},
verbose=verbose,
normalize=normalize)
nspd_kernel.fit(train)
assert is_picklable(nspd_kernel)
gk.fit(train)
assert is_picklable(gk)
def test_neighborhood_subgraph_pairwise_distance():
"""Random input test for the Neighborhood Subgraph Pairwise Distance kernel [+ generic-wrapper]."""
train, test = generate_dataset(n_graphs=100,
r_vertices=(5, 10),
r_connectivity=(0.4, 0.8),
r_weight_edges=(1, 1),
n_graphs_test=40,
random_state=rs,
features=('nl', 5, 'el', 4))
nspd_kernel = NeighborhoodSubgraphPairwiseDistance(verbose=verbose,
normalize=normalize)
gk = GraphKernel(kernel="NSPD", verbose=verbose, normalize=normalize)
try:
nspd_kernel.fit_transform(train)
nspd_kernel.transform(test)
gk.fit_transform(train)
gk.transform(test)
assert True
except Exception as exception:
assert False, exception
lambda: HadamardCode(
n_iter=1, n_jobs=N_JOBS, normalize=NORMALIZING_GRAPH_KERNELS),
"GK-HC-2":
lambda: HadamardCode(
n_iter=2, n_jobs=N_JOBS, normalize=NORMALIZING_GRAPH_KERNELS),
"GK-HC-3":
lambda: HadamardCode(
n_iter=3, n_jobs=N_JOBS, normalize=NORMALIZING_GRAPH_KERNELS),
"GK-HC-4":
lambda: HadamardCode(
n_iter=4, n_jobs=N_JOBS, normalize=NORMALIZING_GRAPH_KERNELS),
"GK-HC-5":
lambda: HadamardCode(
n_iter=5, n_jobs=N_JOBS, normalize=NORMALIZING_GRAPH_KERNELS),
"GK-NSPD":
lambda: NeighborhoodSubgraphPairwiseDistance(normalize=
NORMALIZING_GRAPH_KERNELS),
"GK-WL-OA-1":
lambda: WeisfeilerLehmanOptimalAssignment(
n_iter=1, n_jobs=N_JOBS, normalize=NORMALIZING_GRAPH_KERNELS),
"GK-WL-OA-2":
lambda: WeisfeilerLehmanOptimalAssignment(
n_iter=2, n_jobs=N_JOBS, normalize=NORMALIZING_GRAPH_KERNELS),
"GK-WL-OA-3":
lambda: WeisfeilerLehmanOptimalAssignment(
n_iter=3, n_jobs=N_JOBS, normalize=NORMALIZING_GRAPH_KERNELS),
"GK-WL-OA-4":
lambda: WeisfeilerLehmanOptimalAssignment(
n_iter=4, n_jobs=N_JOBS, normalize=NORMALIZING_GRAPH_KERNELS),
"GK-WL-OA-5":
lambda: WeisfeilerLehmanOptimalAssignment(
n_iter=5, n_jobs=N_JOBS, normalize=NORMALIZING_GRAPH_KERNELS),
for i in range(n_samples):
node_labels = dict()
for j in range(height):
for k in range(width):
node_labels[j * height + k] = int(images[i, j, k] / 4)
graphs.append(
Graph(edges, node_labels=node_labels, edge_labels=edge_labels))
print("Splitting dataset into train/test (1000/100 instances)\n")
graphs_train, graphs_test = graphs[:1000], graphs[1000:1100]
y_train, y_test = y[:1000], y[1000:1100]
# Initialize neighborhood subgraph pairwise distance kernel
gk = NeighborhoodSubgraphPairwiseDistance(r=3, d=2)
print("Computing kernel matrics\n")
t0 = time.time()
K_train = gk.fit_transform(graphs_train)
K_test = gk.transform(graphs_test)
print("done in %0.3fs\n" % (time.time() - t0))
print("Classifying digits\n")
# Initialize SVM
clf = SVC(kernel='precomputed')
# Fit on the train Kernel
clf.fit(K_train, y_train)
# Predict and test.