def _simulate(self):
# Note: the features matrix already exists, and is created by
# the super class
features = self.features
n_samples, n_features = features.shape
link = self.link
u = features.dot(self.weights)
# Add the intercept if necessary
if self.intercept is not None:
u += self.intercept
# Compute the intensities
if link == "identity":
if np.any(u <= 0):
raise ValueError(("features and weights leads to ,",
"non-negative intensities for ", "Poisson."))
intensity = u
else:
intensity = np.exp(u)
# Simulate the Poisson variables. We want it in float64 for
# later computations, hence the next line.
labels = np.empty(n_samples)
labels[:] = poisson(intensity)
self._set("labels", labels)
return features, labels
def poisson(mean, shape=[]):
"""poisson(mean) or poisson(mean, [n, m, ...]) returns array of poisson
distributed random integers with specified mean."""
if shape == []:
shape = None
return mt.poisson(mean, shape)
def poisson(mean, shape=[]):
"""poisson(mean) or poisson(mean, [n, m, ...]) returns array of poisson
distributed random integers with specified mean."""
if shape == []:
shape = None
return mt.poisson(mean, shape)
def test_scatter(self):
"""gnuplot.scatter test (interactive only)"""
from numpy.random.mtrand import poisson
if self.local:
self.p = scatter(poisson(50, (1000, 2)))
def test_scatter(self):
"""gnuplot.scatter test (interactive only)"""
from numpy.random.mtrand import poisson
if self.local:
self.p = scatter( poisson(50,(1000,2)) )