def testParameters(self):
pdf = Dirichlet([1, 1])
alpha = numpy.array([2, 2])
pdf.alpha = numpy.array([2, 2])
self.assertEqual(pdf.alpha.tolist(), alpha.tolist())
self.assertEqual(pdf.alpha.tolist(), pdf['alpha'].tolist())
def testParameters(self):
pdf = Dirichlet([1, 1])
alpha = numpy.array([2, 2])
pdf.alpha = numpy.array([2, 2])
self.assertEqual(pdf.alpha.tolist(), alpha.tolist())
self.assertEqual(pdf.alpha.tolist(), pdf['alpha'].tolist())
def testParameterEstimation(self):
alpha = numpy.array([1., 1., 1.])
pdf = Dirichlet(alpha)
self.assertAlmostEqual(pdf(numpy.array([1., 0., 0])),
pdf(numpy.array([0., 1., 0])),
delta=1e-1)
self.assertAlmostEqual(pdf(numpy.array([1. / 3., 1. / 3., 1. / 3])),
pdf(numpy.array([0., 1., 0])),
delta=1e-1)
alpha = numpy.array([16., 16., 16.])
pdf = Dirichlet(alpha)
self.assertAlmostEqual(pdf(numpy.array([1., 0., 0])),
pdf(numpy.array([0., 1., 0])),
delta=1e-1)
self.assertAlmostEqual(pdf(numpy.array([1., 0., 0])),
pdf(numpy.array([0., 0., 1])),
delta=1e-1)
self.assertNotEqual(pdf(numpy.array([1. / 3., 1. / 3., 1. / 3])),
pdf(numpy.array([0., 1., 0])))
alpha = numpy.array([1., 1., 1.])
pdf = Dirichlet(alpha)
samples = pdf.random(100000)
pdf.estimate(samples)
for i in range(len(alpha)):
self.assertAlmostEqual(pdf.alpha[i], alpha[i], delta=1e-1)
alpha = numpy.array([200., 60., 20.])
pdf = Dirichlet(alpha)
samples = pdf.random(100000)
pdf.estimate(samples)
for i in range(len(alpha)):
self.assertAlmostEqual(pdf.alpha[i], alpha[i], delta=1e1)
def testLogProb(self):
alpha = numpy.array([2, 2])
pdf = Dirichlet(alpha)
x = numpy.array([0.5, 0.5])
self.assertAlmostEqual(pdf(x), 1.5, places=2)
x = numpy.array([0.1, 0.9])
self.assertAlmostEqual(pdf(x), .54, places=2)
x = numpy.array([0.9, 0.1])
self.assertAlmostEqual(pdf(x), .54, places=2)
x = numpy.array([[0.9, 0.1], [0.1, 0.9], [0.5, 0.5]])
self.assertAlmostEqual(pdf(x)[0], .54, places=2)
self.assertAlmostEqual(pdf(x)[1], .54, places=2)
self.assertAlmostEqual(pdf(x)[2], 1.5, places=2)
def testParameterEstimation(self):
alpha = numpy.array([1., 1., 1.])
pdf = Dirichlet(alpha)
self.assertAlmostEqual(pdf(numpy.array([1., 0., 0])),
pdf(numpy.array([0., 1., 0])),
delta=1e-1)
self.assertAlmostEqual(pdf(numpy.array([1. / 3., 1. / 3., 1. / 3])),
pdf(numpy.array([0., 1., 0])),
delta=1e-1)
alpha = numpy.array([16., 16., 16.])
pdf = Dirichlet(alpha)
self.assertAlmostEqual(pdf(numpy.array([1., 0., 0])),
pdf(numpy.array([0., 1., 0])),
delta=1e-1)
self.assertAlmostEqual(pdf(numpy.array([1., 0., 0])),
pdf(numpy.array([0., 0., 1])),
delta=1e-1)
self.assertNotEqual(pdf(numpy.array([1. / 3., 1. / 3., 1. / 3])),
pdf(numpy.array([0., 1., 0])))
alpha = numpy.array([1., 1., 1.])
pdf = Dirichlet(alpha)
samples = pdf.random(100000)
pdf.estimate(samples)
for i in range(len(alpha)):
self.assertAlmostEqual(pdf.alpha[i], alpha[i], delta=1e-1)
alpha = numpy.array([200., 60., 20.])
pdf = Dirichlet(alpha)
samples = pdf.random(100000)
pdf.estimate(samples)
for i in range(len(alpha)):
self.assertAlmostEqual(pdf.alpha[i], alpha[i], delta=1e1)
