vSMC
vSMC: Scalable Monte Carlo
|
Multivariate Normal random walk proposal. More...
#include <vsmc/rng/internal/common.hpp>
Public Types | |
using | result_type = RealType |
Public Member Functions | |
NormalMVProposal (const result_type *chol=nullptr, const result_type *a=nullptr, const result_type *b=nullptr) | |
Only usable when Dim != Dynamic More... | |
NormalMVProposal (std::size_t dim, const result_type *chol=nullptr, const result_type *a=nullptr, const result_type *b=nullptr) | |
const result_type * | a () const |
const result_type * | b () const |
std::size_t | dim () const |
template<typename RNGType > | |
result_type | operator() (RNGType &rng, std::size_t, const result_type *x, result_type *y) |
Multivariate Normal random walk proposal.
Definition at line 543 of file common.hpp.
using vsmc::NormalMVProposal< RealType, Dim >::result_type = RealType |
Definition at line 470 of file random_walk.hpp.
|
inlineexplicit |
Only usable when Dim != Dynamic
chol | The lower triangular elements of the Cholesky decomposition of the covaraince matrix, packed row by row. If it is a nullpointer, then the covariance is the identicy matrix \(I\) |
a | The lower bound of the support of the target distribuiton. It is assumed that the support is \(\prod_{p=1}^d E_p \subset \mathbb{R}^d\) where \(E_p \subset \mathbb{R}\). |
b | The upper bound of the support of the target distribution. |
If the geometry of the support is more complex than above, then one may find a superset of the support that takes the required form, and reject proposals that lay outside the support manually.
Definition at line 486 of file random_walk.hpp.
|
inlineexplicit |
Definition at line 495 of file random_walk.hpp.
|
inline |
Definition at line 506 of file random_walk.hpp.
|
inline |
Definition at line 507 of file random_walk.hpp.
|
inline |
Definition at line 505 of file random_walk.hpp.
|
inline |
Definition at line 510 of file random_walk.hpp.