def test_bernoulli(self): bernoulli = Bernoulli(.725) self.assertAlmostEqual(mean(bernoulli.sample(1000000)), .725, 2) samples = bernoulli.sample(10) loglik = log(.725) * samples + log(0.275) * (1. - samples) self.assertLess(max(abs(loglik - bernoulli.loglikelihood(samples))), 1e-8)
def test_bernoulli_pickle(self):
tmp_file = mkstemp()[1]
p0 = Bernoulli(.3)
with open(tmp_file, 'w') as handle:
dump({'p': p0}, handle)
with open(tmp_file) as handle:
p1 = load(handle)['p']
x = p0.sample(100)
self.assertLess(max(abs(p0.loglikelihood(x) - p1.loglikelihood(x))), 1e-6)
def test_glm_basics(self):
glm = GLM(4, LogisticFunction, Bernoulli)
x = randn(1000)
f = glm.nonlinearity
y = f(x).ravel()
for i in range(x.size):
self.assertAlmostEqual(y[i], 1. / (1. + exp(-x[i])))
glm.nonlinearity = f
y = glm.nonlinearity(x).ravel()
for i in range(x.size):
self.assertAlmostEqual(y[i], 1. / (1. + exp(-x[i])))
b = Bernoulli()
glm = GLM(4, f, b)
glm.nonlinearity = f
y = glm.nonlinearity(x).ravel()
for i in range(x.size):
self.assertAlmostEqual(y[i], 1. / (1. + exp(-x[i])))
self.assertTrue(isinstance(glm.distribution, Bernoulli))
# test wrong order of arguments
self.assertRaises(TypeError,
lambda: GLM(5, Bernoulli, LogisticFunction))
