def test_poisson_reg_params():
"""Test whether the upper bound on the regularization parameters correctly
zero out the coefficients."""
n_features = 5
n_samples = 1000
X, y, beta, intercept = make_poisson_regression(
n_samples=n_samples,
n_features=n_features,
n_informative=n_features,
random_state=2332)
alpha_sets = [np.array([0.5]), np.array([1.0])]
for alpha_set in alpha_sets:
uoi_poisson = UoI_Poisson(alphas=alpha_set)
reg_params = uoi_poisson.get_reg_params(X, y)
alpha = reg_params[0]['alpha']
l1_ratio = reg_params[0]['l1_ratio']
# check that coefficients get set to zero
poisson = Poisson(alpha=1.01 * alpha,
l1_ratio=l1_ratio,
standardize=False,
fit_intercept=True)
poisson.fit(X, y)
assert_equal(poisson.coef_, 0.)
# check that coefficients below the bound are not set to zero
poisson = Poisson(alpha=0.99 * alpha,
l1_ratio=l1_ratio,
standardize=False,
fit_intercept=True)
poisson.fit(X, y)
assert np.count_nonzero(poisson.coef_) > 0
def test_fit_intercept():
"""Tests whether `include_intercept` in passed through to the linear models.
"""
poi = UoI_Poisson(fit_intercept=True)
assert poi._selection_lm.fit_intercept
assert poi._estimation_lm.fit_intercept
poi = UoI_Poisson(fit_intercept=False)
assert not poi._selection_lm.fit_intercept
assert not poi._estimation_lm.fit_intercept
def test_poisson():
"""Test basic functionality of UoI_Lasso and that it finds right model"""
n_features = 20
n_inf = 10
X, y, w, b = make_poisson_regression(n_samples=200,
n_features=n_features,
n_informative=n_inf,
random_state=10)
poisson = UoI_Poisson(comm=MPI.COMM_WORLD)
poisson.fit(X, y)
assert (np.sign(abs(w)) == np.sign(abs(poisson.coef_))).mean() >= .6
def test_UoI_Poisson_all_zeros():
"""Test that UoI Poisson correctly fits the data when the response variable
is all zeros."""
n_features = 5
n_samples = 100
X = np.random.normal(size=(n_samples, n_features))
y = np.zeros(n_samples)
poisson = UoI_Poisson()
poisson.fit(X, y)
assert_equal(poisson.intercept_, -np.inf)
assert_equal(poisson.coef_, np.zeros(n_features))
def test_score_predictions():
"""Test the score predictions function in UoI Poisson."""
X = np.array([[np.log(2), -1, -3], [np.log(3), -2, -4],
[np.log(4), -3, -5], [np.log(5), -4, -6]])
y = 1. / np.log([2., 3., 4., 5.])
support = np.array([True, False, False])
# create fitter by hand
fitter = Poisson()
fitter.coef_ = np.array([1])
fitter.intercept_ = 0
# test log-likelihood
ll = UoI_Poisson.score_predictions(metric='log',
fitter=fitter,
X=X,
y=y,
support=support)
assert_almost_equal(ll, -2.5)
# test information criteria
aic = UoI_Poisson.score_predictions(metric='AIC',
fitter=fitter,
X=X,
y=y,
support=support)
assert_almost_equal(aic, 2 * ll - 2)
aicc = UoI_Poisson.score_predictions(metric='AICc',
fitter=fitter,
X=X,
y=y,
support=support)
assert_almost_equal(aicc, aic - 2)
bic = UoI_Poisson.score_predictions(metric='BIC',
fitter=fitter,
X=X,
y=y,
support=support)
assert_almost_equal(bic, 2 * ll - np.log(y.size))
# test invalid metric
assert_raises(ValueError, UoI_Poisson.score_predictions, 'fake', fitter, X,
y, support)
def test_UoI_Poisson():
"""Tests the UoI Poisson fitter with lbfgs solver."""
n_features = 3
n_samples = 5000
# create data
X, y, beta, _ = make_poisson_regression(n_samples=n_samples,
n_features=n_features,
n_informative=n_features,
beta=np.array([0., 1.0, 1.5]),
random_state=2332)
# lbfgs
poisson = UoI_Poisson(n_lambdas=48, n_boots_sel=30, n_boots_est=30,
alphas=np.array([1.0]), warm_start=False)
poisson.fit(X, y)
assert_allclose(poisson.coef_, beta, atol=0.5)
def test_estimation_target():
"""Verify the ability for the user to set the estimation taget variable"""
# Assess r2 on train data
uoi = UoI_Lasso(estimation_score='r2', estimation_target='train')
# train gets converted to the index 0
assert (uoi._estimation_target == 0)
# Assess BIC on test data
uoi = UoI_Lasso(estimation_score='BIC', estimation_target='test')
# Assess r2 on train data
uoi = UoI_ElasticNet(estimation_score='r2', estimation_target='train')
# train gets converted to the index 0
assert (uoi._estimation_target == 0)
# Assess BIC on test data
uoi = UoI_ElasticNet(estimation_score='BIC', estimation_target='test')
assert (uoi._estimation_target == 1)
uoi = UoI_L1Logistic(estimation_score='acc', estimation_target='train')
assert (uoi._estimation_target == 0)
uoi = UoI_L1Logistic(estimation_score='BIC', estimation_target='test')
assert (uoi._estimation_target == 1)
uoi = UoI_Poisson(estimation_score='acc', estimation_target='train')
assert (uoi._estimation_target == 0)
uoi = UoI_Poisson(estimation_score='BIC', estimation_target='test')
assert (uoi._estimation_target == 1)