def al_stream(self, data, target, X_train, y_train, X_full, y_full, train_idx):
# initializing the active learner
acc = []
learner = ActiveLearner(
estimator=RandomForestClassifier(),
query_strategy=margin_sampling,
X_training=X_train, y_training=y_train
)
# print('Initial prediction accuracy: %f' % learner.score(X_full, y_full))
index = 0
# learning until the accuracy reaches a given threshold
xx = 0;
while learner.score(X_full, y_full) < 0.90:
print(xx)
xx = xx + 1
stream_idx = np.random.choice(range(len(X_full)))
if classifier_entropy(learner, X_full[stream_idx].reshape(1, -1)) >= 0.2:
learner.teach(X_full[stream_idx].reshape(1, -1), y_full[stream_idx].reshape(-1, ))
learner_score = learner.score(X_full, y_full)
# print('Item no. %d queried, new accuracy: %f' % (stream_idx, learner_score))
# print('%0.3f' % (learner_score), end=",")
if index == self.query_number:
break
index = index + 1
acc.append(learner_score)
return acc
def rob_sampler(classifier, X, n_instances=1, **uncertainty_measure_kwargs):
"""
Sample randomly with sample weights based on entropy
:param classifier:
:param X:
:param n_instances:
:param uncertainty_measure_kwargs:
:return:
"""
entropy = classifier_entropy(classifier, X, **uncertainty_measure_kwargs)
sample_weights = entropy / np.sum(entropy)
query_idx = np.random.choice(range(len(entropy)),
size=n_instances,
p=sample_weights)
return query_idx, X[query_idx]