ldaplusplus::em::CorrespondenceSupervisedMStep< Scalar > Class Template Reference

#include <CorrespondenceSupervisedMStep.hpp>

Inheritance diagram for ldaplusplus::em::CorrespondenceSupervisedMStep< Scalar >:

Public Member Functions
	CorrespondenceSupervisedMStep (Scalar mu=2.)
virtual void	m_step (std::shared_ptr< parameters::Parameters > parameters) override
virtual void	doc_m_step (const std::shared_ptr< corpus::Document > doc, const std::shared_ptr< parameters::Parameters > v_parameters, std::shared_ptr< parameters::Parameters > m_parameters) override
Public Member Functions inherited from ldaplusplus::events::EventDispatcherComposition
std::shared_ptr< EventDispatcherInterface >	get_event_dispatcher ()
void	set_event_dispatcher (std::shared_ptr< EventDispatcherInterface > dispatcher)

Detailed Description

template<typename Scalar>
class ldaplusplus::em::CorrespondenceSupervisedMStep< Scalar >

CorrespondenceSupervisedMStep implements the maximization step of a variant of the correspondence LDA model.

This model implements the correspondence LDA but instead of labels it tries to also generate the class of the document. The generative procedure is the following:

1. Given \(K\) \(V\)-dimensional multinomials as the topics and \(K\) \(C\)-dimensional multinomials to sample class labels from
2. Sample from a Dirichlet the topic mixture from a document
3. For N times
   1. Sample a topic \(k\)
   2. From that topic sample a word using the \(k\)-th \(V\)-dimensional multinomial
4. Sample from a uniform distribution a number \(n\) between 1 and N
5. Using the previously sampled \(n\)-th topic sample a class label from the multinomials

Member Function Documentation

template<typename Scalar >

void ldaplusplus::CorrespondenceSupervisedMStep::doc_m_step ( const std::shared_ptr< corpus::Document >  doc, const std::shared_ptr< parameters::Parameters >  v_parameters, std::shared_ptr< parameters::Parameters >  m_parameters  )

overridevirtual

Count the occurences of every word and class to implement maximum likelihood estimation in the m_step()

Parameters

doc	A single document
v_parameters	The variational parameters used in m-step in order to maximize model parameters
m_parameters	Model parameters, used as output in case of online methods

Implements ldaplusplus::em::MStepInterface< Scalar >.

template<typename Scalar >

void ldaplusplus::CorrespondenceSupervisedMStep::m_step ( std::shared_ptr< parameters::Parameters >  parameters )

overridevirtual

Maximize the ELBO w.r.t. to \(\beta\) and \(\eta\) (the topics and class generating multinomials).

Parameters

parameters

Model parameters (maybe changed after call)

Implements ldaplusplus::em::MStepInterface< Scalar >.

---

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

• include/ldaplusplus/em/CorrespondenceSupervisedMStep.hpp
• src/ldaplusplus/em/CorrespondenceSupervisedMStep.cpp