def compute_scores(adj_matrix, labels, n_folds, n_bagging_models, sample_ratio,
n_kfold_trials):
clf = TransductiveBaggingClassifier(1, -1, n_bagging_models, sample_ratio,
n_jobs)
predicted_vs_true_tuples = []
for i in range(n_kfold_trials):
for train_idxs, test_idxs in KFold(labels.shape[0],
n_folds=n_folds,
shuffle=True):
predict_labels = labels.copy()
predict_labels[test_idxs] = -1
clf.fit(adj_matrix, predict_labels)
predicted = clf.transduction_
predicted_vs_true_tuples.append(
(predicted[test_idxs], labels[test_idxs]))
# print predicted[test_idxs]
# print labels[test_idxs]
all_scores = []
for label_index in range(labels.shape[1]):
scores = []
for pred_labels, true_labels in predicted_vs_true_tuples:
try:
scores.append(
metrics.roc_auc_score(true_labels[:, label_index],
pred_labels[:, label_index]))
except ValueError, e:
print("Error computing roc_auc_score, label={}: {}".format(
label_index, e))
all_scores.append(scores)
def compute_scores(adj_matrix, labels, n_folds, n_bagging_models,
sample_ratio, n_kfold_trials):
clf = TransductiveBaggingClassifier(1, -1, n_bagging_models,
sample_ratio, n_jobs)
predicted_vs_true_tuples = []
for i in range(n_kfold_trials):
for train_idxs, test_idxs in KFold(
labels.shape[0],
n_folds=n_folds,
shuffle=True):
predict_labels = labels.copy()
predict_labels[test_idxs] = -1
clf.fit(adj_matrix, predict_labels)
predicted = clf.transduction_
predicted_vs_true_tuples.append(
(predicted[test_idxs], labels[test_idxs]))
# print predicted[test_idxs]
# print labels[test_idxs]
all_scores = []
for label_index in range(labels.shape[1]):
scores = []
for pred_labels, true_labels in predicted_vs_true_tuples:
try:
scores.append(metrics.roc_auc_score(
true_labels[:, label_index],
pred_labels[:, label_index]))
except ValueError, e:
print("Error computing roc_auc_score, label={}: {}".format(label_index, e))
all_scores.append(scores)
def test_transduction_bagging2(self):
clf = TransductiveBaggingClassifier(1, 1000, 6, 0.5)
g = nx.Graph()
## two squares w/ vertices joined
g.add_edges_from([(0, 1), (0, 2), (0, 3), (1, 2), (1, 3), (2, 3),
(3, 4),
(4, 5), (4, 6), (4, 7), (5, 6), (5, 7), (6, 7)])
labels = np.array([-1, 0, 0, 0, 1, 1, 1, -1], ndmin=2).T
adjacency_matrix = nx.to_numpy_matrix(g).A
clf.fit(adjacency_matrix, labels)
np.testing.assert_array_equal(
clf.transduction_,
np.array([0,0,0,0,1,1,1,1], ndmin=2).T)
def test_transduction_bagging(self):
clf = TransductiveBaggingClassifier(1, 1000, 6, 0.5)
g = nx.Graph()
# graph of three cpts, 0, 1, 2 and 3, 4, 5
g.add_edges_from([(0, 1), (1, 2), (0, 2), (3, 4), (4, 5)])
labels = np.array([[0, 0, -1, 1, 1, -1],
[1, 1, -1, 0, 0, -1]]).T
adjacency_matrix = nx.to_numpy_matrix(g).A
clf.fit(adjacency_matrix, labels)
np.testing.assert_array_equal(
clf.transduction_,
np.array([[0, 0, 0, 1, 1, 1],
[1, 1, 1, 0, 0, 0]]).T)