191 return String(
_T(
"unknown exception"));
321#ifdef STK_MIXTURE_DEBUG
323 {
stk_cout <<
_T(
"In stringToMixture, mixture ") << type <<
_T(
" not found.\n");}
442#ifdef STK_MIXTURE_DEBUG
443 stk_cout <<
_T(
"In stringToMixture, mixture ") << type <<
_T(
" not found.\n");
692 p_algo->setNbIterMax(nbIterMax);
693 p_algo->setEpsilon(epsilon);
719 p_algo->setNbIterMax(nbIterMax);
720 p_algo->setEpsilon(epsilon);
746 p_algo->setNbIterBurn(nbIterBurn);
748 p_algo->setEpsilon(epsilon);
In this file we define the abstract base class for mixture models.
In this file we define learning mixture algorithms.
In this file we define algorithms for predicting in a mixture model.
In this file we define mixture algorithms.
In this file we define the classes for computing the penalized criterion on mixture models.
In this file we define the interface base class for initialization methods.
In this file we define the strategies to use in order to estimate a mixture model.
#define stk_cout
Standard stk output stream.
#define _T(x)
Let x unmodified.
Derived class of Criterion for computing the AIC Criterion The AIC criteria of a given model M is a p...
Derived class of Criterion for computing the BIC Criterion The Bic criteria of a model M is a penaliz...
Implementation of the SEM algorithm.
Initialization by simulating a realization of the class labels zi accordingly to the initial proporti...
Implementation of the EM algorithm.
Implementation of the EMPredict algorithm.
A FullStrategy is based on the following paradigm:
Initialization by simulating the tik accordingly to the initial proportions.
Derived class of IMixtureCriterion for computing the ICL Criterion The ICL criteria of a model M is a...
Interface base class for the learning algorithms.
Interface base class for predicting algorithms.
Interface base class for the algorithms.
Base class for Mixture (composed) model.
Interface base class for the selection model criterion.
Interface base class for the initializations.
Interface base class for all the strategies.
Implementation of the ImputeAlgo learning algorithm.
Derived class of Criterion for computing the Maximum Likelihood Criterion This criterion does not pen...
The MultidimRegression class allows to regress a multidimensional output variable among a multivariat...
Implementation of the random initialization.
Implementation of the SEM algorithm.
Implementation of the SemiSEM algorithm.
Implementation of the SemiSEMPredict algorithm.
A SimpleStrategy is just nbTry long run.
Implementation of the SimulAlgo learning algorithm.
String const & toUpperString(String &s)
convert the characters of the String to upper case
algoPredictType
Learning estimation algorithms.
initType
initialization type.
algoType
Estimation algorithms.
MixtureClass
list of the class of mixture implemented in stkpp
algoType stringToAlgo(String const &type)
Convert a String to an algoType.
initType stringToInit(String const &type)
Convert a String to a initType.
MixtureClass mixtureToMixtureClass(Mixture const &type)
convert a Mixture to a MixtureClass.
algoLearnType
Learning estimation algorithms.
IMixtureInit * createInit(Clust::initType init=defaultInitType, int nbInits=defaultNbInit, Clust::algoType algo=defaultAlgoInInit, int nbIterMax=defaultNbIterMaxInInit, Real epsilon=defaultEpsilonInInit)
Utility function for creating a model initializer.
IMixtureAlgo * createAlgo(Clust::algoType algo, int nbIterMax, Real epsilon)
utility function for creating an estimation algorithm.
algoPredictType stringToPredictAlgo(String const &type)
Convert a String to an algoPredictType.
IMixtureStrategy * createSimpleStrategy(IMixtureComposer *&p_composer, int nbTry, IMixtureInit *const &p_init, IMixtureAlgo *const &algo)
Utility function for creating a SimpleStrategy.
IMixtureCriterion * createCriterion(Clust::criterionType criterion)
criterionType stringToCriterion(String const &type)
Convert a String to an criterionType.
Mixture stringToMixture(String const &type)
Convert a String to a Mixture.
String mixtureToString(Mixture const &type)
convert a Mixture to a String.
String exceptionToString(exceptions const &type)
convert a Clust::exceptions to a String.
IMixtureAlgoLearn * createLearnAlgo(Clust::algoLearnType algo, int nbIterMax, Real epsilon)
utility function for creating a learning algorithm.
algoLearnType stringToLearnAlgo(String const &type)
Convert a String to an algoLearnType.
exceptions
Specific exceptions allowing to handle the erroros that can occur in the estimation process.
Mixture
list of the mixtures that can be used by the composer
IMixtureStrategy * createFullStrategy(IMixtureComposer *&p_composer, int nbTry, int nbInitRun, IMixtureInit *const &p_init, int nbShortRun, IMixtureAlgo *const &shortRunAlgo, IMixtureAlgo *const &longRunAlgo)
Utility function for creating a FullStrategy.
IMixtureAlgoPredict * createPredictAlgo(Clust::algoPredictType algo, int nbIterBurn, int nbIterLong, Real epsilon)
utility function for creating a predicting algorithm.
criterionType
type of criterion to use in order to select the mixture model
@ randomParamInit_
initialize randomly the parameters
@ randomFuzzyInit_
initialize randomly the partnership class probabilities
@ randomClassInit_
initialize randomly the class labels
@ noInit_
no initialization
std::basic_string< Char > String
STK fundamental type of a String.
double Real
STK fundamental type of Real values.
The namespace STK is the main domain space of the Statistical ToolKit project.
helper structure encapsulating the parameters of the Full strategy
helper structure encapsulating the parameters of the simple strategy