How to use the abcpy.distributions.MultiStudentT function in abcpy

To help you get started, we’ve selected a few abcpy examples, based on popular ways it is used in public projects.

Secure your code as it's written. Use Snyk Code to scan source code in minutes - no build needed - and fix issues immediately.

eth-cscs / abcpy / tests / distributions_tests.py View on Github

def test_set_parameters(self):
        #Test initialization and setting of parameters with fixed values
        distribution = MultiStudentT([1.2,1.3],[[1,0],[0,1]],2,seed=1)
        distribution.set_parameters([[1.5,1.6],[[1.1,0],[0,1.1]],3])
        self.assertTrue(distribution.get_parameters()==[[1.5, 1.6], [[1.1, 0], [0, 1.1]], 3])

        #Test initialization and setting of parameters with 1 multidimensional distribution
        helper_distribution = MultiStudentT([1,1],[[0.1,0],[0,0.1]],2)
        distribution = MultiStudentT(helper_distribution, [[1,0],[0,1]],3,seed=1)
        distribution.set_parameters([[[[1.2,1.3],[[1.2,1.4],[[1,0],[0,1]],2]]],[[1.1,0],[0,1.2]],2])
        self.assertTrue(distribution.get_parameters()==[[[[1.2,1.3],[[1.2,1.4],[[1,0],[0,1]],2]]],[[1.1,0],[0,1.2]],2])

        #Test initialization and setting of parameters with 2 distributions
        helper_distribution = Normal(1,0.5,seed=1)
        helper_distribution_2 = Normal(2,0.1,seed=1)
        distribution = MultiStudentT([helper_distribution,helper_distribution_2],[[1,0],[0,1]],2)
        distribution.set_parameters([[[1.1,[1.2,0.3]],[2.2,[2.1,0.2]]],[[1.1,0],[0,1.1]],3])
        self.assertTrue(distribution.get_parameters()==[[[1.1,[1.2,0.3]],[2.2,[2.1,0.2]]],[[1.1,0],[0,1.1]],3])

        #Test initialization and setting of parameters with distribution for degrees of freedom
        distribution = MultiStudentT([1,1],[[1,0],[0,1]],helper_distribution)
        distribution.set_parameters([[1.1,1.2],[[0.1,0],[0,0.1]],[1,[1.4,0.7]]])
        self.assertTrue(distribution.get_parameters()==[[1.1,1.2],[[0.1,0],[0,0.1]],[1,[1.4,0.7]]])

eth-cscs / abcpy / tests / distributions_tests.py View on Github

def test_set_parameters(self):
        #Test initialization and setting of parameters with fixed values
        distribution = MultiStudentT([1.2,1.3],[[1,0],[0,1]],2,seed=1)
        distribution.set_parameters([[1.5,1.6],[[1.1,0],[0,1.1]],3])
        self.assertTrue(distribution.get_parameters()==[[1.5, 1.6], [[1.1, 0], [0, 1.1]], 3])

        #Test initialization and setting of parameters with 1 multidimensional distribution
        helper_distribution = MultiStudentT([1,1],[[0.1,0],[0,0.1]],2)
        distribution = MultiStudentT(helper_distribution, [[1,0],[0,1]],3,seed=1)
        distribution.set_parameters([[[[1.2,1.3],[[1.2,1.4],[[1,0],[0,1]],2]]],[[1.1,0],[0,1.2]],2])
        self.assertTrue(distribution.get_parameters()==[[[[1.2,1.3],[[1.2,1.4],[[1,0],[0,1]],2]]],[[1.1,0],[0,1.2]],2])

        #Test initialization and setting of parameters with 2 distributions
        helper_distribution = Normal(1,0.5,seed=1)
        helper_distribution_2 = Normal(2,0.1,seed=1)
        distribution = MultiStudentT([helper_distribution,helper_distribution_2],[[1,0],[0,1]],2)
        distribution.set_parameters([[[1.1,[1.2,0.3]],[2.2,[2.1,0.2]]],[[1.1,0],[0,1.1]],3])
        self.assertTrue(distribution.get_parameters()==[[[1.1,[1.2,0.3]],[2.2,[2.1,0.2]]],[[1.1,0],[0,1.1]],3])

        #Test initialization and setting of parameters with distribution for degrees of freedom
        distribution = MultiStudentT([1,1],[[1,0],[0,1]],helper_distribution)
        distribution.set_parameters([[1.1,1.2],[[0.1,0],[0,0.1]],[1,[1.4,0.7]]])
        self.assertTrue(distribution.get_parameters()==[[1.1,1.2],[[0.1,0],[0,0.1]],[1,[1.4,0.7]]])

eth-cscs / abcpy / tests / distributions_tests.py View on Github

def test_set_parameters(self):
        #Test initialization and setting of parameters with fixed values
        distribution = MultiStudentT([1.2,1.3],[[1,0],[0,1]],2,seed=1)
        distribution.set_parameters([[1.5,1.6],[[1.1,0],[0,1.1]],3])
        self.assertTrue(distribution.get_parameters()==[[1.5, 1.6], [[1.1, 0], [0, 1.1]], 3])

        #Test initialization and setting of parameters with 1 multidimensional distribution
        helper_distribution = MultiStudentT([1,1],[[0.1,0],[0,0.1]],2)
        distribution = MultiStudentT(helper_distribution, [[1,0],[0,1]],3,seed=1)
        distribution.set_parameters([[[[1.2,1.3],[[1.2,1.4],[[1,0],[0,1]],2]]],[[1.1,0],[0,1.2]],2])
        self.assertTrue(distribution.get_parameters()==[[[[1.2,1.3],[[1.2,1.4],[[1,0],[0,1]],2]]],[[1.1,0],[0,1.2]],2])

        #Test initialization and setting of parameters with 2 distributions
        helper_distribution = Normal(1,0.5,seed=1)
        helper_distribution_2 = Normal(2,0.1,seed=1)
        distribution = MultiStudentT([helper_distribution,helper_distribution_2],[[1,0],[0,1]],2)
        distribution.set_parameters([[[1.1,[1.2,0.3]],[2.2,[2.1,0.2]]],[[1.1,0],[0,1.1]],3])
        self.assertTrue(distribution.get_parameters()==[[[1.1,[1.2,0.3]],[2.2,[2.1,0.2]]],[[1.1,0],[0,1.1]],3])

eth-cscs / abcpy / tests / distributions_tests.py View on Github

def test_set_parameters(self):
        #Test initialization and setting of parameters with fixed values
        distribution = MultiStudentT([1.2,1.3],[[1,0],[0,1]],2,seed=1)
        distribution.set_parameters([[1.5,1.6],[[1.1,0],[0,1.1]],3])
        self.assertTrue(distribution.get_parameters()==[[1.5, 1.6], [[1.1, 0], [0, 1.1]], 3])

        #Test initialization and setting of parameters with 1 multidimensional distribution
        helper_distribution = MultiStudentT([1,1],[[0.1,0],[0,0.1]],2)
        distribution = MultiStudentT(helper_distribution, [[1,0],[0,1]],3,seed=1)
        distribution.set_parameters([[[[1.2,1.3],[[1.2,1.4],[[1,0],[0,1]],2]]],[[1.1,0],[0,1.2]],2])
        self.assertTrue(distribution.get_parameters()==[[[[1.2,1.3],[[1.2,1.4],[[1,0],[0,1]],2]]],[[1.1,0],[0,1.2]],2])

        #Test initialization and setting of parameters with 2 distributions
        helper_distribution = Normal(1,0.5,seed=1)
        helper_distribution_2 = Normal(2,0.1,seed=1)
        distribution = MultiStudentT([helper_distribution,helper_distribution_2],[[1,0],[0,1]],2)
        distribution.set_parameters([[[1.1,[1.2,0.3]],[2.2,[2.1,0.2]]],[[1.1,0],[0,1.1]],3])
        self.assertTrue(distribution.get_parameters()==[[[1.1,[1.2,0.3]],[2.2,[2.1,0.2]]],[[1.1,0],[0,1.1]],3])

        #Test initialization and setting of parameters with distribution for degrees of freedom
        distribution = MultiStudentT([1,1],[[1,0],[0,1]],helper_distribution)
        distribution.set_parameters([[1.1,1.2],[[0.1,0],[0,0.1]],[1,[1.4,0.7]]])
        self.assertTrue(distribution.get_parameters()==[[1.1,1.2],[[0.1,0],[0,0.1]],[1,[1.4,0.7]]])

eth-cscs / abcpy / tests / distributions_tests.py View on Github

def test_sample(self):
        m = np.array([0, 0])
        cov = np.eye(2)
        distribution = MultiStudentT(m, cov, 4)
        samples = distribution.sample(10000000)
        expected_mean = np.array([0., 0.])
        expected_var = np.array([1.9978, 2.0005])
        diff_mean = np.abs(samples.mean(axis=0) - expected_mean)
        diff_var = np.abs(samples.var(axis=0) - expected_var)
        self.assertLess(diff_mean.sum(), 2e-2)
        self.assertLess(diff_var.sum(), 2e-2)

        distribution = MultiStudentT(distribution, cov, 4, seed=1)
        samples_graph = distribution.sample(10000000)
        expected_mean_graph = np.array([0.006, 0.003])
        expected_var_graph = np.array([1.99683813, 2.00319229])
        diff_mean_graph = np.abs(samples_graph.mean(axis=0) - expected_mean_graph)

        diff_var_graph = np.abs(samples_graph.var(axis=0)-expected_var_graph)
        self.assertLess(diff_mean_graph.sum(), 3.5) #this value doesnt work at aaaaalllllll

eth-cscs / abcpy / tests / distributions_tests.py View on Github

def test_sample(self):
        m = np.array([0, 0])
        cov = np.eye(2)
        distribution = MultiStudentT(m, cov, 4)
        samples = distribution.sample(10000000)
        expected_mean = np.array([0., 0.])
        expected_var = np.array([1.9978, 2.0005])
        diff_mean = np.abs(samples.mean(axis=0) - expected_mean)
        diff_var = np.abs(samples.var(axis=0) - expected_var)
        self.assertLess(diff_mean.sum(), 2e-2)
        self.assertLess(diff_var.sum(), 2e-2)

        distribution = MultiStudentT(distribution, cov, 4, seed=1)
        samples_graph = distribution.sample(10000000)
        expected_mean_graph = np.array([0.006, 0.003])
        expected_var_graph = np.array([1.99683813, 2.00319229])
        diff_mean_graph = np.abs(samples_graph.mean(axis=0) - expected_mean_graph)

        diff_var_graph = np.abs(samples_graph.var(axis=0)-expected_var_graph)
        self.assertLess(diff_mean_graph.sum(), 3.5) #this value doesnt work at aaaaalllllll
        self.assertLess(diff_var_graph.sum(), 2e-2)

        distribution = MultiStudentT(Uniform([1,2],[3,4],seed=1),cov,2,seed=1)
        distribution = MultiStudentT([Normal(1,0.5,1),StudentT(1,2,seed=1)],cov,2,seed=1)

eth-cscs / abcpy / tests / distributions_tests.py View on Github

def test_pdf(self):
        m = [0, 0]
        cov = [[1,0],[0,1]]
        distribution = MultiStudentT(m, cov, 1)
        self.assertLess(abs(distribution.pdf([0., 0.]) - 0.15915), 1e-5)

        cov = np.array([[2, 0], [0, 2]])
        distribution = MultiStudentT(m, cov, 1)
        self.assertLess(abs(distribution.pdf([0., 0.]) - 0.079577), 1e-5)
        self.assertLess(abs(distribution.pdf([1., 1.]) - 0.028135), 1e-5)

eth-cscs / abcpy / examples / backends / apache_spark / pmcabc_gaussian.py View on Github

# define the model
    from abcpy.models import Gaussian
    model = Gaussian(prior, seed=1)

    # define statistics
    from abcpy.statistics import Identity
    statistics_calculator = Identity(degree = 2, cross = False)

    # define distance
    from abcpy.distances import LogReg
    distance_calculator = LogReg(statistics_calculator)

    # define kernel
    from abcpy.distributions import MultiStudentT
    mean, cov, df = np.array([.0, .0]), np.eye(2), 3.
    kernel = MultiStudentT(mean, cov, df, seed=1)


    # define sampling scheme
    from abcpy.inferences import PMCABC
    sampler = PMCABC(model, distance_calculator, kernel, backend, seed=1)
    
    # sample from scheme
    T, n_sample, n_samples_per_param = 3, 250, 10
    eps_arr = np.array([.75])
    epsilon_percentile = 10
    journal = sampler.sample(y_obs,  T, eps_arr, n_sample, n_samples_per_param, epsilon_percentile)

    return journal

eth-cscs / abcpy / examples / extensions / simplemodels / gaussian_f90 / pmcabc-gaussian_model_simple.py View on Github

# define the model
    model = Gaussian(prior)
    
    # define statistics
    from abcpy.statistics import Identity
    statistics_calculator = Identity(degree = 2, cross = False)
    
    # define distance
    from abcpy.distances import LogReg
    distance_calculator = LogReg(statistics_calculator)
    
    # define kernel
    from abcpy.distributions import MultiStudentT
    mean, cov, df = np.array([.0, .0]), np.eye(2), 3.
    kernel = MultiStudentT(mean, cov, df)
    
    # define backend
    from abcpy.backends import BackendSpark as Backend
    from abcpy.backends import BackendDummy as Backend
    backend = Backend()
    
    # define sampling scheme
    from abcpy.inferences import PMCABC
    sampler = PMCABC(model, distance_calculator, kernel, backend)
    
    # sample from scheme
    T, n_sample, n_samples_per_param = 3, 100, 10
    eps_arr = np.array([.75])
    epsilon_percentile = 10
    journal = sampler.sample(y_obs,  T, eps_arr, n_sample, n_samples_per_param, epsilon_percentile)

How to use the abcpy.distributions.MultiStudentT function in abcpy

To help you get started, we’ve selected a few abcpy examples, based on popular ways it is used in public projects.

abcpy

Package Health Score

Popular abcpy functions

Similar packages