def test_methods(N, p, seed, opt):
X, y = make_intercept_data(N, p, seed=seed)
coefs = opt(X, y)
p = sigmoid(X.dot(coefs).compute())
y_sum = y.compute().sum()
p_sum = p.sum()
assert np.isclose(y_sum, p_sum, atol=1e-1)
def broadcast_lbfgs_weight():
with cluster() as (s, [a, b]):
with Client(s['address'], loop=loop) as c:
X, y = make_intercept_data(1000, 10)
coefs = lbfgs(X, y, dask_distributed_client=c)
p = sigmoid(X.dot(coefs).compute())
y_sum = y.compute().sum()
p_sum = p.sum()
assert np.isclose(y_sum, p_sum, atol=1e-1)
def predict_proba(self, X):
"""Probability estimates for samples in X.
Parameters
----------
X : array-like, shape = [n_samples, n_features]
Returns
-------
T : array-like, shape = [n_samples, n_classes]
The probability of the sample for each class in the model.
"""
return sigmoid(self.decision_function(X))
def test_methods(N, p, seed, opt, is_cupy):
X, y = make_intercept_data(N, p, seed=seed)
if is_cupy:
cupy = pytest.importorskip('cupy')
X, y = to_dask_cupy_array_xy(X, y, cupy)
coefs = opt(X, y)
p = sigmoid(X.dot(coefs).compute())
y_sum = y.compute().sum()
p_sum = p.sum()
assert np.isclose(y_sum, p_sum, atol=1e-1)
def predict_proba(self, X):
"""Probability estimates for samples in X.
Parameters
----------
X : array-like, shape = [n_samples, n_features]
Returns
-------
T : array-like, shape = [n_samples, n_classes]
The probability of the sample for each class in the model.
"""
X_ = self._check_array(X)
return sigmoid(dot(X_, self._coef))
def predict_proba(self, X):
"""Probability estimates for samples in X.
Parameters
----------
X : array-like, shape = [n_samples, n_features]
Returns
-------
T : array-like, shape = [n_samples, n_classes]
The probability of the sample for each class in the model.
"""
# TODO: more work needed here to support multi_class
prob = sigmoid(self.decision_function(X))
return lr_prob_stack(prob)
def hessian(self, Xbeta, X):
"""Logistic hessian"""
p = sigmoid(Xbeta)
return dot(p * (1 - p) * X.T, X)
def gradient(self, Xbeta, X, y):
"""Logistic gradient"""
p = sigmoid(Xbeta)
return dot(X.T, p - y)
def hessian(Xbeta, X):
p = sigmoid(Xbeta)
return dot(p * (1 - p) * X.T, X)
def gradient(Xbeta, X, y):
p = sigmoid(Xbeta)
return dot(X.T, p - y)
def predict_proba(self, X):
from dask_glm.utils import sigmoid
X = self._input_to_dask_cupy_array(X)
return sigmoid(self.decision_function(X))
