def test_regression_scorers():
# Test regression scorers.
diabetes = load_diabetes()
X, y = diabetes.data, diabetes.target
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0)
clf = Ridge()
clf.fit(X_train, y_train)
score1 = get_scorer('r2')(clf, X_test, y_test)
score2 = r2_score(y_test, clf.predict(X_test))
assert_almost_equal(score1, score2)
def test_sag_regressor_computed_correctly():
"""tests if the sag regressor is computed correctly"""
alpha = .1
n_features = 10
n_samples = 40
max_iter = 50
tol = .000001
fit_intercept = True
rng = np.random.RandomState(0)
X = rng.normal(size=(n_samples, n_features))
w = rng.normal(size=n_features)
y = np.dot(X, w) + 2.
step_size = get_step_size(X, alpha, fit_intercept, classification=False)
clf1 = Ridge(fit_intercept=fit_intercept,
tol=tol,
solver='sag',
alpha=alpha * n_samples,
max_iter=max_iter)
clf2 = clone(clf1)
clf1.fit(X, y)
clf2.fit(sp.csr_matrix(X), y)
spweights1, spintercept1 = sag_sparse(X,
y,
step_size,
alpha,
n_iter=max_iter,
dloss=squared_dloss,
fit_intercept=fit_intercept)
spweights2, spintercept2 = sag_sparse(X,
y,
step_size,
alpha,
n_iter=max_iter,
dloss=squared_dloss,
sparse=True,
fit_intercept=fit_intercept)
assert_array_almost_equal(clf1.coef_.ravel(),
spweights1.ravel(),
decimal=3)
assert_almost_equal(clf1.intercept_, spintercept1, decimal=1)
def test_sag_regressor():
"""tests if the sag regressor performs well"""
xmin, xmax = -5, 5
n_samples = 20
tol = .001
max_iter = 50
alpha = 0.1
rng = np.random.RandomState(0)
X = np.linspace(xmin, xmax, n_samples).reshape(n_samples, 1)
# simple linear function without noise
y = 0.5 * X.ravel()
clf1 = Ridge(tol=tol,
solver='sag',
max_iter=max_iter,
alpha=alpha * n_samples,
random_state=rng)
clf2 = clone(clf1)
clf1.fit(X, y)
clf2.fit(sp.csr_matrix(X), y)
score1 = clf1.score(X, y)
score2 = clf2.score(X, y)
assert score1 > 0.99
assert score2 > 0.99
# simple linear function with noise
y = 0.5 * X.ravel() + rng.randn(n_samples, 1).ravel()
clf1 = Ridge(tol=tol,
solver='sag',
max_iter=max_iter,
alpha=alpha * n_samples)
clf2 = clone(clf1)
clf1.fit(X, y)
clf2.fit(sp.csr_matrix(X), y)
score1 = clf1.score(X, y)
score2 = clf2.score(X, y)
score2 = clf2.score(X, y)
assert score1 > 0.5
assert score2 > 0.5
def test_huber_better_r2_score():
# Test that huber returns a better r2 score than non-outliers"""
X, y = make_regression_with_outliers()
huber = HuberRegressor(alpha=0.01)
huber.fit(X, y)
linear_loss = np.dot(X, huber.coef_) + huber.intercept_ - y
mask = np.abs(linear_loss) < huber.epsilon * huber.scale_
huber_score = huber.score(X[mask], y[mask])
huber_outlier_score = huber.score(X[~mask], y[~mask])
# The Ridge regressor should be influenced by the outliers and hence
# give a worse score on the non-outliers as compared to the huber
# regressor.
ridge = Ridge(alpha=0.01)
ridge.fit(X, y)
ridge_score = ridge.score(X[mask], y[mask])
ridge_outlier_score = ridge.score(X[~mask], y[~mask])
assert huber_score > ridge_score
# The huber model should also fit poorly on the outliers.
assert ridge_outlier_score > huber_outlier_score
def test_regressor_matching():
n_samples = 10
n_features = 5
rng = np.random.RandomState(10)
X = rng.normal(size=(n_samples, n_features))
true_w = rng.normal(size=n_features)
y = X.dot(true_w)
alpha = 1.
n_iter = 100
fit_intercept = True
step_size = get_step_size(X, alpha, fit_intercept, classification=False)
clf = Ridge(fit_intercept=fit_intercept,
tol=.00000000001,
solver='sag',
alpha=alpha * n_samples,
max_iter=n_iter)
clf.fit(X, y)
weights1, intercept1 = sag_sparse(X,
y,
step_size,
alpha,
n_iter=n_iter,
dloss=squared_dloss,
fit_intercept=fit_intercept)
weights2, intercept2 = sag(X,
y,
step_size,
alpha,
n_iter=n_iter,
dloss=squared_dloss,
fit_intercept=fit_intercept)
assert_allclose(weights1, clf.coef_)
assert_allclose(intercept1, clf.intercept_)
assert_allclose(weights2, clf.coef_)
assert_allclose(intercept2, clf.intercept_)
def test_sag_pobj_matches_ridge_regression():
"""tests if the sag pobj matches ridge reg"""
n_samples = 100
n_features = 10
alpha = 1.0
n_iter = 100
fit_intercept = False
rng = np.random.RandomState(10)
X = rng.normal(size=(n_samples, n_features))
true_w = rng.normal(size=n_features)
y = X.dot(true_w)
clf1 = Ridge(fit_intercept=fit_intercept,
tol=.00000000001,
solver='sag',
alpha=alpha,
max_iter=n_iter,
random_state=42)
clf2 = clone(clf1)
clf3 = Ridge(fit_intercept=fit_intercept,
tol=.00001,
solver='lsqr',
alpha=alpha,
max_iter=n_iter,
random_state=42)
clf1.fit(X, y)
clf2.fit(sp.csr_matrix(X), y)
clf3.fit(X, y)
pobj1 = get_pobj(clf1.coef_, alpha, X, y, squared_loss)
pobj2 = get_pobj(clf2.coef_, alpha, X, y, squared_loss)
pobj3 = get_pobj(clf3.coef_, alpha, X, y, squared_loss)
assert_array_almost_equal(pobj1, pobj2, decimal=4)
assert_array_almost_equal(pobj1, pobj3, decimal=4)
assert_array_almost_equal(pobj3, pobj2, decimal=4)
X, y, w = make_regression(n_samples=10,
n_features=10,
coef=True,
random_state=1,
bias=3.5)
coefs = []
errors = []
alphas = np.logspace(-6, 6, 200)
# Train the model with different regularisation strengths
for a in alphas:
clf.set_params(alpha=a)
clf.fit(X, y)
coefs.append(clf.coef_)
errors.append(mean_squared_error(clf.coef_, w))
# Display results
plt.figure(figsize=(20, 6))
plt.subplot(121)
ax = plt.gca()
ax.plot(alphas, coefs)
ax.set_xscale('log')
plt.xlabel('alpha')
plt.ylabel('weights')
plt.title('Ridge coefficients as a function of the regularization')
plt.axis('tight')
X_outliers[2:, :] += X.min() - X.mean() / 4.
y_outliers[:2] += y.min() - y.mean() / 4.
y_outliers[2:] += y.max() + y.mean() / 4.
X = np.vstack((X, X_outliers))
y = np.concatenate((y, y_outliers))
plt.plot(X, y, 'b.')
# Fit the huber regressor over a series of epsilon values.
colors = ['r-', 'b-', 'y-', 'm-']
x = np.linspace(X.min(), X.max(), 7)
epsilon_values = [1.35, 1.5, 1.75, 1.9]
for k, epsilon in enumerate(epsilon_values):
huber = HuberRegressor(alpha=0.0, epsilon=epsilon)
huber.fit(X, y)
coef_ = huber.coef_ * x + huber.intercept_
plt.plot(x, coef_, colors[k], label="huber loss, %s" % epsilon)
# Fit a ridge regressor to compare it to huber regressor.
ridge = Ridge(alpha=0.0, random_state=0, normalize=True)
ridge.fit(X, y)
coef_ridge = ridge.coef_
coef_ = ridge.coef_ * x + ridge.intercept_
plt.plot(x, coef_, 'g-', label="ridge regression")
plt.title("Comparison of HuberRegressor vs Ridge")
plt.xlabel("X")
plt.ylabel("y")
plt.legend(loc=0)
plt.show()
mask[(points[0]).astype(np.int), (points[1]).astype(np.int)] = 1
mask = ndimage.gaussian_filter(mask, sigma=l / n_pts)
res = np.logical_and(mask > mask.mean(), mask_outer)
return np.logical_xor(res, ndimage.binary_erosion(res))
# Generate synthetic images, and projections
l = 128
proj_operator = build_projection_operator(l, l // 7)
data = generate_synthetic_data()
proj = proj_operator * data.ravel()[:, np.newaxis]
proj += 0.15 * np.random.randn(*proj.shape)
# Reconstruction with L2 (Ridge) penalization
rgr_ridge = Ridge(alpha=0.2)
rgr_ridge.fit(proj_operator, proj.ravel())
rec_l2 = rgr_ridge.coef_.reshape(l, l)
# Reconstruction with L1 (Lasso) penalization
# the best value of alpha was determined using cross validation
# with LassoCV
rgr_lasso = Lasso(alpha=0.001)
rgr_lasso.fit(proj_operator, proj.ravel())
rec_l1 = rgr_lasso.coef_.reshape(l, l)
plt.figure(figsize=(8, 3.3))
plt.subplot(131)
plt.imshow(data, cmap=plt.cm.gray, interpolation='nearest')
plt.axis('off')
plt.title('original image')
plt.subplot(132)