Conditional< TPDF, TParamMap > Class Template Reference

Conditional distribution function. More...

Public Member Functions
	Conditional (TPDF pdf, TParamMap map)
	Constructors returns a conditional distribution object. More...
template<typename... Args>
auto	random (const Args &...args) -> decltype(std::declval< TPDF >().random())
	Sample from distribution. More...
template<typename... Args>
double	likelihood (const decltype(std::declval< TPDF >().random())&rv, const Args &...args)
	Calculate the likelihood of a random variable. More...
const TPDF &	getPDF () const
	Returns a reference to \(\mathcal{F}(\theta)\). More...
const TParamMap &	getParamMap () const
	Returns a reference to \(g(.)\). More...

Detailed Description

template<typename TPDF, typename TParamMap>
class ssmkit::distribution::Conditional< TPDF, TParamMap >

Conditional distribution function.

Let \(x \in X\) and \((y_0, \cdots, y_N) \in \{Y_0, \cdots, Y_n\} \), the class defines a conditional distribution of \(x\) given \(y_0, \cdots, y_N\):

\begin{equation}p(x|y_0, \cdots, y_N) = \mathcal{F}(g(y_0, \cdots, y_N))\end{equation}

where \(\mathcal{F}(\theta)\) is a probability distribution with parameter \(\theta \in \Theta \) and \(g:\{Y_0, \cdots, Y_n\} \rightarrow \Theta\) is a function that maps a set of variables \({y_0^*, \cdots, y_N^*}\) to a parameter \(\theta^*\), i.e. \(\theta^* = g(y_0^*, \cdots, y_N^*)\).

Template Parameters

TPDF	Type of the distribution \(\mathcal{F}(\theta)\)
TParamMap	Type of the parameter map \(g(.)\)

Bayesian network representation of conditional distribution

Definition at line 34 of file conditional.hpp.

Constructor & Destructor Documentation

Conditional ( TPDF  pdf, TParamMap  map  )

inline

Constructors returns a conditional distribution object.

Note

use makeConditional builder for convenient template argument deduction.

Parameters

pdf	A probability distribution object that implements \(\mathcal{F}(\theta)\).
map	A callable object that implements \(g(.)\).

Precondition

pdf should provide random, likelihood and parameterize methods, e.g. Gaussian. map should be callable with return type equivalent to parameter type of pdf, e.g. map::LinearGaussian.

Definition at line 45 of file conditional.hpp.

Member Function Documentation

auto random ( const Args &...  args ) -> decltype(std::declval<TPDF>().random())

inline

Sample from distribution.

Parameters

args...

Condition variables \(y_0, \cdots, y_N\).

Returns

random variable \(x\).

Definition at line 53 of file conditional.hpp.

double likelihood ( const decltype(std::declval< TPDF >().random())&  rv, const Args &...  args  )

inline

Calculate the likelihood of a random variable.

Parameters

args...	Condition variables \(y_0, \cdots, y_N\).
rv	random variable \(x\).

Returns

likelihood \(p(x|y_0, \cdots, y_N)\).

Definition at line 65 of file conditional.hpp.

const TPDF& getPDF ( ) const

inline

Returns a reference to \(\mathcal{F}(\theta)\).

Definition at line 72 of file conditional.hpp.

const TParamMap& getParamMap ( ) const

inline

Returns a reference to \(g(.)\).

Definition at line 74 of file conditional.hpp.

---

The documentation for this class was generated from the following file:

• src/ssmkit/distribution/conditional.hpp