tfd_von_mises: The von Mises distribution over angles

View source: R/distributions.R

tfd_von_misesR Documentation

The von Mises distribution over angles

Description

The von Mises distribution is a univariate directional distribution. Similarly to Normal distribution, it is a maximum entropy distribution. The samples of this distribution are angles, measured in radians. They are 2 pi-periodic: x = 0 and x = 2pi are equivalent. This means that the density is also 2 pi-periodic. The generated samples, however, are guaranteed to be in [-pi, pi) range. When concentration = 0, this distribution becomes a Uniform distribuion on the [-pi, pi) domain.

Usage

tfd_von_mises(
  loc,
  concentration,
  validate_args = FALSE,
  allow_nan_stats = TRUE,
  name = "VonMises"
)

Arguments

loc

Floating point tensor, the circular means of the distribution(s).

concentration

Floating point tensor, the level of concentration of the distribution(s) around loc. Must take non-negative values. concentration = 0 defines a Uniform distribution, while concentration = +inf indicates a Deterministic distribution at loc.

validate_args

Logical, default FALSE. When TRUE distribution parameters are checked for validity despite possibly degrading runtime performance. When FALSE invalid inputs may silently render incorrect outputs. Default value: FALSE.

allow_nan_stats

Logical, default TRUE. When TRUE, statistics (e.g., mean, mode, variance) use the value NaN to indicate the result is undefined. When FALSE, an exception is raised if one or more of the statistic's batch members are undefined.

name

name prefixed to Ops created by this class.

Details

The von Mises distribution is a special case of von Mises-Fisher distribution for n=2. However, the TFP's VonMisesFisher implementation represents the samples and location as (x, y) points on a circle, while VonMises represents them as scalar angles.

Mathematical details The probability density function (pdf) of this distribution is,

pdf(x; loc, concentration) = exp(concentration cos(x - loc)) / Z
Z = 2 * pi * I_0 (concentration)

where:

  • I_0 (concentration) is the modified Bessel function of order zero;

  • loc the circular mean of the distribution, a scalar. It can take arbitrary values, but it is 2pi-periodic: loc and loc + 2pi result in the same distribution.

  • concentration >= 0 parameter is the concentration parameter. When concentration = 0, this distribution becomes a Uniform distribution on [-pi, pi).

The parameters loc and concentration must be shaped in a way that supports broadcasting (e.g. loc + concentration is a valid operation).

Value

a distribution instance.

See Also

For usage examples see e.g. tfd_sample(), tfd_log_prob(), tfd_mean().

Other distributions: tfd_autoregressive(), tfd_batch_reshape(), tfd_bates(), tfd_bernoulli(), tfd_beta_binomial(), tfd_beta(), tfd_binomial(), tfd_categorical(), tfd_cauchy(), tfd_chi2(), tfd_chi(), tfd_cholesky_lkj(), tfd_continuous_bernoulli(), tfd_deterministic(), tfd_dirichlet_multinomial(), tfd_dirichlet(), tfd_empirical(), tfd_exp_gamma(), tfd_exp_inverse_gamma(), tfd_exponential(), tfd_gamma_gamma(), tfd_gamma(), tfd_gaussian_process_regression_model(), tfd_gaussian_process(), tfd_generalized_normal(), tfd_geometric(), tfd_gumbel(), tfd_half_cauchy(), tfd_half_normal(), tfd_hidden_markov_model(), tfd_horseshoe(), tfd_independent(), tfd_inverse_gamma(), tfd_inverse_gaussian(), tfd_johnson_s_u(), tfd_joint_distribution_named_auto_batched(), tfd_joint_distribution_named(), tfd_joint_distribution_sequential_auto_batched(), tfd_joint_distribution_sequential(), tfd_kumaraswamy(), tfd_laplace(), tfd_linear_gaussian_state_space_model(), tfd_lkj(), tfd_log_logistic(), tfd_log_normal(), tfd_logistic(), tfd_mixture_same_family(), tfd_mixture(), tfd_multinomial(), tfd_multivariate_normal_diag_plus_low_rank(), tfd_multivariate_normal_diag(), tfd_multivariate_normal_full_covariance(), tfd_multivariate_normal_linear_operator(), tfd_multivariate_normal_tri_l(), tfd_multivariate_student_t_linear_operator(), tfd_negative_binomial(), tfd_normal(), tfd_one_hot_categorical(), tfd_pareto(), tfd_pixel_cnn(), tfd_poisson_log_normal_quadrature_compound(), tfd_poisson(), tfd_power_spherical(), tfd_probit_bernoulli(), tfd_quantized(), tfd_relaxed_bernoulli(), tfd_relaxed_one_hot_categorical(), tfd_sample_distribution(), tfd_sinh_arcsinh(), tfd_skellam(), tfd_spherical_uniform(), tfd_student_t_process(), tfd_student_t(), tfd_transformed_distribution(), tfd_triangular(), tfd_truncated_cauchy(), tfd_truncated_normal(), tfd_uniform(), tfd_variational_gaussian_process(), tfd_vector_diffeomixture(), tfd_vector_exponential_diag(), tfd_vector_exponential_linear_operator(), tfd_vector_laplace_diag(), tfd_vector_laplace_linear_operator(), tfd_vector_sinh_arcsinh_diag(), tfd_von_mises_fisher(), tfd_weibull(), tfd_wishart_linear_operator(), tfd_wishart_tri_l(), tfd_wishart(), tfd_zipf()


tfprobability documentation built on Sept. 1, 2022, 5:07 p.m.