def test_random_walk():
"""Picklability test for the Simple Random Walk kernel."""
train, _ = generate_dataset(n_graphs=100,
r_vertices=(10, 20),
r_connectivity=(0.4, 0.8),
r_weight_edges=(0.01, 12.0),
n_graphs_test=40,
random_state=rs,
features=None)
rw_kernel = RandomWalk(verbose=verbose, normalize=normalize)
rw_kernel.fit(train)
assert is_picklable(rw_kernel)
def test_random_walk():
"""Eigenvalue test for the Simple, Labelled Random Walk kernel."""
rw_kernel = RandomWalk(verbose=verbose, normalize=normalize, lamda=0.01)
if verbose:
print_kernel("Random Walk", rw_kernel, dataset_tr, dataset_te)
else:
positive_eig(rw_kernel, dataset)
rw_kernel_lab = RandomWalkLabeled(verbose=verbose, normalize=normalize, lamda=0.0001)
if verbose:
print_kernel("Random Walk Labelled", rw_kernel_lab, dataset_tr, dataset_te)
else:
positive_eig(rw_kernel_lab, dataset)
def test_random_walk():
"""Random input test for the Simple Random Walk kernel."""
train, test = generate_dataset(n_graphs=100,
r_vertices=(10, 20),
r_connectivity=(0.4, 0.8),
r_weight_edges=(0.01, 12.0),
n_graphs_test=40,
random_state=rs,
features=None)
rw_kernel = RandomWalk(verbose=verbose, normalize=normalize)
try:
rw_kernel.fit_transform(train)
rw_kernel.transform(test)
assert True
except Exception as exception:
assert False, exception
"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),
}
NOT_TESTED = {
"GK-ShortestPathA":
lambda: ShortestPathAttr(normalize=NORMALIZING_GRAPH_KERNELS),
"GK-RandomWalk":
lambda: RandomWalk(n_jobs=N_JOBS, normalize=NORMALIZING_GRAPH_KERNELS
), # taking too long
"GK-RandomWalkLabeled":
lambda: RandomWalkLabeled(
n_jobs=N_JOBS, normalize=NORMALIZING_GRAPH_KERNELS), # taking too long
"GK-GraphHopper":
lambda: GraphHopper(normalize=NORMALIZING_GRAPH_KERNELS),
"GK-PyramidMatch":
lambda: PyramidMatch(normalize=NORMALIZING_GRAPH_KERNELS), # Error with PG
"GK-LovaszTheta":
lambda: LovaszTheta(normalize=NORMALIZING_GRAPH_KERNELS),
"GK-SvmTheta":
lambda: SvmTheta(normalize=NORMALIZING_GRAPH_KERNELS),
"GK-Propagation":
lambda: Propagation(normalize=NORMALIZING_GRAPH_KERNELS),
"GK-PropagationA":
lambda: PropagationAttr(normalize=NORMALIZING_GRAPH_KERNELS),
K_train = gk.fit_transform(G_train)
K_test = gk.transform(G_test)
clf = SVC(kernel='precomputed', C=1) # Initialize SVM
clf.fit(K_train, y_train) # Train SVM
y_pred = clf.predict(K_test) # Predict
# Compute the classification accuracy
# hint: use the accuracy_score function of scikit-learn
print("Classification accuracy using ShortestPath",
accuracy_score(y_test, y_pred))
# Use the random walk kernel and the pyramid match graph kernel to perform classification
gk = RandomWalk()
K_train = gk.fit_transform(G_train)
K_test = gk.transform(G_test)
clf = SVC(kernel='precomputed', C=1) # Initialize SVM
clf.fit(K_train, y_train) # Train SVM
y_pred = clf.predict(K_test) # Predict
print("Classification accuracy using RandomWalk",
accuracy_score(y_test, y_pred))
gk = PyramidMatch(with_labels=False)
K_train = gk.fit_transform(G_train)
K_test = gk.transform(G_test)
clf.fit(K_train, y_train) # Train SVM y_pred = clf.predict(K_test) # Predict # Compute the classification accuracy # hint: use the accuracy_score function of scikit-learn ################## # your code here # print(accuracy_score(y_test, y_pred)) ################## # Use the random walk kernel and the pyramid match graph kernel to perform classification ################## # your code here # gk1 = RandomWalk() K_train1 = gk1.fit_transform(G_train) K_test1 = gk1.transform(G_test) clf1 = SVC(kernel='precomputed', C=1) # Initialize SVM clf1.fit(K_train, y_train) # Train SVM y_pred1 = clf1.predict(K_test) # Predict print(accuracy_score(y_test, y_pred1)) ################## ############## Question 3 # Classify the graphs of a real-world dataset using graph kernels # Load the MUTAG dataset # hint: use the fetch_dataset function of GraKeL