The record of a HiddenMarkovFilter's forward pass that a smoother runs backward over: one CategoricalPredictedJoint per prediction, each with room for its smoothed distribution. More...

#include <EmrysFilter/Smoothers/CategoricalHistory.h>

Classes
struct	Step
	One recorded prediction and its smoothed distribution. More...

Public Types
using	Scalar = Scalar_
	the scalar type
using	SpaceType = Space_
	the discrete space
using	CategoricalType = CategoricalParams<Space_, Scalar_>
	the smoothed distribution's type
using	JointType = CategoricalPredictedJoint<Space_, Scalar_>
	what each step records
using	allocator_type = Allocator_
	the allocator the steps come from
using	Index = Eigen::Index
	Eigen's index type.
using	const_iterator = typename ListType::const_iterator
	iterates the steps, oldest first

Public Member Functions
	CategoricalHistory ()=default
	An empty history that keeps every step.
	CategoricalHistory (Allocator_ const &allocator)
	An empty history that keeps every step, with nodes from allocator.
	CategoricalHistory (Lag lag, Allocator_ const &allocator={})
	An empty history that keeps the newest lag steps, with nodes from allocator.
	CategoricalHistory (CategoricalHistory const &other)
	Copies the steps into nodes from other's allocator (see Allocation).
	CategoricalHistory (CategoricalHistory const &other, Allocator_ const &allocator)
	Copies other's steps into nodes from allocator.
	CategoricalHistory (CategoricalHistory &&other)=default
	Takes other's steps.
	CategoricalHistory (CategoricalHistory &&other, Allocator_ const &allocator)
	Moves other's steps into nodes from allocator, taking its nodes when the allocators are equal.
CategoricalHistory &	operator= (CategoricalHistory const &other)=default
	Copies other's steps and lag.
CategoricalHistory &	operator= (CategoricalHistory &&other)=default
	Takes other's steps and lag, as `std::list's` move assignment does.
allocator_type	get_allocator () const
std::optional< Index >	lag () const
Index	size () const
bool	empty () const
Step const &	front () const
Step const &	back () const
const_iterator	begin () const
const_iterator	end () const
void	push (JointType joint)
	Records joint as the newest step.
void	smooth (CategoricalType const &latest)
	Fills every step's smoothed distribution by smoothStep(), from the newest back, given latest, the filter's state after the newest step: its posterior \( p(x_N \mid z_{1:N}) \), which is its own smoothed distribution.
void	clear ()
	Drops every step, keeping the lag.

Detailed Description

template<DiscreteSpace Space_, RealScalar Scalar_ = double, typename Allocator_ = DefaultAllocator>
class Emrys::CategoricalHistory< Space_, Scalar_, Allocator_ >

The record of a HiddenMarkovFilter's forward pass that a smoother runs backward over: one CategoricalPredictedJoint per prediction, each with room for its smoothed distribution.

GaussianHistory for a discrete state (decisions.md [smoothing], [emissions]), with the same interface: push() records the joint a filter's predictJoint() returns, smooth() takes the filter's current state and fills every step's smoothed distribution by smoothStep(), from the newest step back, and a Lag makes it a fixed-lag window that recycles its oldest node. Updates between predictions need no recording. The same record gives the most probable sequence (mostProbableSequence()) and the pairwise transition posteriors (transitionPosterior()).

CategoricalHistory<Weather> history;
for (auto const & z : observations)
{
    history.push(filter.predictJoint());
    static_cast<void>(filter.update(table.observe(z)));
}
history.smooth(filter.state());
for (auto const & step : history) { use(step.smoothed); }       // p(x_k | z_{1:N}), oldest first
auto const path = mostProbableSequence(history, filter.state());  // Viterbi

For a bounded space every step is inline in its node, so a fixed-lag history allocates nothing once full.

Template Parameters

Space_	the discrete space
Scalar_	scalar type of the probabilities, `double` by default
Allocator_	allocator for the steps, re-bound to the list's nodes (see Allocation)

See also: smoothStep(), CategoricalPredictedJoint, HiddenMarkovFilter::predictJoint(), Smoothers