laplace_marginal_density.hpp File Reference

Detailed Description

Reference for calculations of marginal and its gradients: Margossian et al (2020), https://arxiv.org/abs/2004.12550 and Margossian (2023), https://arxiv.org/pdf/2306.14976.

Definition in file laplace_marginal_density.hpp.

#include <stan/math/prim/fun/Eigen.hpp>
#include <stan/math/mix/functor/laplace_likelihood.hpp>
#include <stan/math/rev/meta.hpp>
#include <stan/math/rev/core.hpp>
#include <stan/math/rev/fun.hpp>
#include <stan/math/rev/fun/value_of.hpp>
#include <stan/math/rev/functor.hpp>
#include <stan/math/prim/fun/to_ref.hpp>
#include <stan/math/prim/fun/quad_form_diag.hpp>
#include <stan/math/prim/functor/iter_tuple_nested.hpp>
#include <unsupported/Eigen/MatrixFunctions>
#include <cmath>
#include <optional>

Go to the source code of this file.

Classes
struct	stan::math::laplace_options_base
	Options for the laplace sampler. More...
struct	stan::math::laplace_options< false >
struct	stan::math::laplace_options< true >
struct	stan::math::internal::laplace_density_estimates< Covar, ThetaVec, WR, L_t, A_vec, ThetaGrad, LU_t, KRoot >

Namespaces
namespace	stan
	The lgamma implementation in stan-math is based on either the reentrant safe lgamma_r implementation from C or the boost::math::lgamma implementation.
namespace	stan::math
	Matrices and templated mathematical functions.
namespace	stan::math::internal
	A comparator that works for any container type that has the brackets operator.

Typedefs
using	stan::math::laplace_options_default = laplace_options< false >
using	stan::math::laplace_options_user_supplied = laplace_options< true >

Functions
template<typename WRootMat >
void	stan::math::internal::block_matrix_sqrt (WRootMat &W_root, const Eigen::SparseMatrix< double > &W, const Eigen::Index block_size)
	Returns the principal square root of a block diagonal matrix.
template<typename WRootMat >
void	stan::math::internal::block_matrix_chol_L (WRootMat &W_root, const Eigen::SparseMatrix< double > &W, const Eigen::Index block_size)
	Performs a Cholesky decomposition on a block diagonal matrix.
template<typename AVec , typename APrev , typename ThetaVec , typename LLFun , typename LLArgs , typename Covar , typename Msgs >
void	stan::math::internal::line_search (double &objective_new, AVec &a, ThetaVec &theta, APrev &a_prev, LLFun &&ll_fun, LLArgs &&ll_args, Covar &&covariance, const int max_steps_line_search, const double objective_old, double tolerance, Msgs *msgs)
	Performs a simple line search.
template<typename Output >
void	stan::math::internal::set_zero_adjoint (Output &&output)
	Set all adjoints of the output to zero.
template<bool ZeroInput = false, typename Output , typename Input , require_t< is_all_arithmetic_scalar< Output > > * = nullptr, require_t< is_all_var_scalar< Input > > * = nullptr>
void	stan::math::internal::collect_adjoints (Output &output, Input &&input)
	Collect the adjoints from the input and add them to the output.
template<typename NameStr , typename ParamStr , typename Param >
void	stan::math::internal::throw_nan (NameStr &&name_str, ParamStr &&param_str, Param &&param)
	Throws an error if the parameter contains NaN or Inf values.
template<typename LLFun , typename LLTupleArgs , typename CovarFun , typename CovarArgs , bool InitTheta, require_t< is_all_arithmetic_scalar< CovarArgs > > * = nullptr>
auto	stan::math::internal::laplace_marginal_density_est (LLFun &&ll_fun, LLTupleArgs &&ll_args, CovarFun &&covariance_function, CovarArgs &&covar_args, const laplace_options< InitTheta > &options, std::ostream *msgs)
	For a latent Gaussian model with hyperparameters phi and latent variables theta, and observations y, this function computes an approximation of the log marginal density, p(y | phi).
template<typename LLFun , typename LLTupleArgs , typename CovarFun , typename CovarArgs , bool InitTheta, require_t< is_all_arithmetic_scalar< CovarArgs, LLTupleArgs > > * = nullptr>
double	stan::math::laplace_marginal_density (LLFun &&ll_fun, LLTupleArgs &&ll_args, CovarFun &&covariance_function, CovarArgs &&covar_args, const laplace_options< InitTheta > &options, std::ostream *msgs)
	For a latent Gaussian model with global parameters phi, latent variables theta, and observations y, this function computes an approximation of the log marginal density, p(y | phi).
template<typename Output , typename Input , require_t< is_all_arithmetic_scalar< Output > > * = nullptr, require_t< is_all_arithmetic_scalar< Input > > * = nullptr>
void	stan::math::internal::collect_adjoints (Output &&output, Input &&input)
	Collects the adjoints from the input and adds them to the output.
template<bool ZeroInput = false>
void constexpr	stan::math::internal::copy_compute_s2 (const std::tuple<> &output, const std::tuple<> &input) noexcept
	Base case for zero sized tuples.
template<bool ZeroInput = false, typename Output , typename Input , require_t< is_all_arithmetic_scalar< Output > > * = nullptr, require_t< is_any_var_scalar< Input > > * = nullptr>
void	stan::math::internal::copy_compute_s2 (Output &&output, Input &&input)
	Copies the adjoints from the input to the output, scaling them by 0.5.
template<typename T >
constexpr decltype(auto)	stan::math::internal::filter_var_scalar_types (T &&t)
template<typename Input >
constexpr auto	stan::math::internal::make_zeroed_arena (Input &&input)
	Creates an arena type from the input with initialized with zeros.
template<typename Output , typename Input >
void	stan::math::internal::collect_adjoints (Output &&output, const vari *ret, Input &&input)
	Used in reverse pass to collect adjoints to the output.
template<typename Output , typename Input >
void	stan::math::internal::reverse_pass_collect_adjoints (var ret, Output &&output, Input &&input)
	Collects adjoints from a tuple or std::vector of tuples.
template<typename LLFun , typename LLTupleArgs , typename CovarFun , typename CovarArgs , bool InitTheta, require_t< is_any_var_scalar< LLTupleArgs, CovarArgs > > * = nullptr>
auto	stan::math::laplace_marginal_density (const LLFun &ll_fun, LLTupleArgs &&ll_args, CovarFun &&covariance_function, CovarArgs &&covar_args, const laplace_options< InitTheta > &options, std::ostream *msgs)
	For a latent Gaussian model with global parameters phi, latent variables theta, and observations y, this function computes an approximation of the log marginal density, p(y | phi).