def test_iris_ensemble_iterative_regression():
print('\ntest_iris_ensemble_iterative_regression():')
X_train, X_test, y_train, y_test = train_test_split(X_iris,
y_iris,
test_size=5,
random_state=42)
cls = ESNClassifier(input_to_node=[('tanh',
InputToNode(hidden_layer_size=10,
random_state=42,
activation='identity')),
('bounded_relu',
InputToNode(hidden_layer_size=10,
random_state=42,
activation='bounded_relu'))
],
node_to_node=[('default',
NodeToNode(hidden_layer_size=20,
spectral_radius=0.0))],
regressor=IncrementalRegression(alpha=.01),
random_state=42)
for samples in np.split(np.arange(0, X_train.shape[0]), 5):
cls.partial_fit(X_train[samples, :],
y_train[samples],
classes=np.arange(3, dtype=int))
y_predicted = cls.predict(X_test)
for record in range(len(y_test)):
print('predicted: {0} \ttrue: {1}'.format(y_predicted[record],
y_test[record]))
print('score: {0}'.format(cls.score(X_test, y_test)))
print('proba: {0}'.format(cls.predict_proba(X_test)))
print('log_proba: {0}'.format(cls.predict_log_proba(X_test)))
assert cls.score(X_test, y_test) >= 4. / 5.
def test_esn_classifier_partial_fit() -> None:
X, y = load_digits(return_X_y=True, as_sequence=True)
esn = ESNClassifier(hidden_layer_size=50, verbose=True)
for k in range(10):
esn.partial_fit(X[k],
np.repeat(y[k], 8),
classes=np.arange(10),
postpone_inverse=True)
print(esn.__sizeof__())
print(esn.hidden_layer_state)
esn = ESNClassifier(hidden_layer_size=50, regressor=Ridge())
with pytest.raises(BaseException):
for k in range(10):
esn.partial_fit(X[k],
np.repeat(y[k], 8),
classes=np.arange(10),
postpone_inverse=True)
# In[ ]:
for params in ParameterGrid(grid):
print(params)
input_to_node = clone(base_input_to_node)
node_to_node = clone(base_node_to_node)
node_to_node.set_params(**params)
esn = ESNClassifier(input_to_node=input_to_node,
node_to_node=node_to_node,
regressor=FastIncrementalRegression(alpha=5e-3),
random_state=10)
for X, y in zip(X_train, y_train):
y = np.repeat(y, repeats=8, axis=0)
esn.partial_fit(X=X, y=y.reshape(-1, 1), classes=range(10))
err_train = []
for X, y in zip(X_train, y_train):
y = np.repeat(np.atleast_2d(y), repeats=8, axis=0)
y_pred = esn.predict(X=X)
err_train.append(zero_one_loss(y, y_pred))
err_test = []
for X, y in zip(X_test, y_test):
y = np.repeat(np.atleast_2d(y), repeats=8, axis=0)
y_pred = esn.predict(X=X)
err_test.append(zero_one_loss(y, y_pred))
print('{0}\t{1}'.format(np.mean(err_train), np.mean(err_test)))
# ## Update parameter of the basic ESN
#
# After optimizing bias and leakage, we update our basic ESN with the identified values for bias and leakage.