regressionTemplate< T > Class Template Reference

#include <regression.h>

Inheritance diagram for regressionTemplate< T >:

Collaboration diagram for regressionTemplate< T >:

Public Member Functions
	regressionTemplate ()
	regressionTemplate (const std::vector< trainingExampleTemplate< T > > &trainingSet)
	regressionTemplate (const int &numInputs, const int &numOutputs)
	~regressionTemplate ()
bool	train (const std::vector< trainingExampleTemplate< T > > &trainingSet) override
float	getTrainingProgress ()
std::vector< size_t >	getNumEpochs () const
void	setNumEpochs (const size_t &epochs)
std::vector< size_t >	getNumHiddenLayers () const
void	setNumHiddenLayers (const int &num_hidden_layers)
std::vector< size_t >	getNumHiddenNodes () const
void	setNumHiddenNodes (const int &num_hidden_nodes)
Public Member Functions inherited from modelSet< T >
	modelSet ()
virtual	~modelSet ()
bool	reset ()
std::vector< T >	run (const std::vector< T > &inputVector)
std::string	getJSON ()
void	writeJSON (const std::string &filepath)
bool	putJSON (const std::string &jsonMessage)
bool	readJSON (const std::string &filepath)

Additional Inherited Members
Protected Member Functions inherited from modelSet< T >
void	threadTrain (std::size_t i, const std::vector< trainingExampleTemplate< T > > &training_set)
Protected Attributes inherited from modelSet< T >
std::vector< baseModel< T > * >	myModelSet
int	numInputs
std::vector< std::string >	inputNames
int	numOutputs
bool	isTraining
bool	isTrained

Detailed Description

template<typename T>
class regressionTemplate< T >

Class for implementing a set of regression models.

This doesn't do anything modelSet can't do. But, it's simpler and more like wekinator. It has some calls that are specifc to neural networks

Constructor & Destructor Documentation

regressionTemplate() [1/3]

template<typename T >

regressionTemplate< T >::regressionTemplate

with no arguments, just make an empty vector

regressionTemplate() [2/3]

template<typename T >

regressionTemplate< T >::regressionTemplate

(

const std::vector< trainingExampleTemplate< T > > &

trainingSet

)

create based on training set inputs and outputs

regressionTemplate() [3/3]

template<typename T >

regressionTemplate< T >::regressionTemplate	(	const int &	numInputs,
		const int &	numOutputs
	)

create with proper models, but not trained

~regressionTemplate()

template<typename T >

regressionTemplate< T >::~regressionTemplate ( )

inline

destructor

Member Function Documentation

getNumEpochs()

template<typename T >

std::vector< size_t > regressionTemplate< T >::getNumEpochs

Check how many training epochs each model will run. This feature is temporary, and will be replaced by a different design.

getNumHiddenLayers()

template<typename T >

std::vector< size_t > regressionTemplate< T >::getNumHiddenLayers

Check how many hidden layers are in each model. This feature is temporary, and will be replaced by a different design.

getNumHiddenNodes()

template<typename T >

std::vector< size_t > regressionTemplate< T >::getNumHiddenNodes

Check how many hidden nodes are in each model. This feature is temporary, and will be replaced by a different design.

getTrainingProgress()

template<typename T >

float regressionTemplate< T >::getTrainingProgress

Check how far the training has gotten. Averages progress over all models in training

setNumEpochs()

template<typename T >

void regressionTemplate< T >::setNumEpochs

(

const size_t &

epochs

)

Call before train, to set the number of training epochs

setNumHiddenLayers()

template<typename T >

void regressionTemplate< T >::setNumHiddenLayers

(

const int &

num_hidden_layers

)

Set how many hidden layers are in all models. This feature is temporary, and will be replaced by a different design.

setNumHiddenNodes()

template<typename T >

void regressionTemplate< T >::setNumHiddenNodes

(

const int &

num_hidden_nodes

)

Set how many hidden layers are in all models. This feature is temporary, and will be replaced by a different design.

train()

template<typename T >

bool regressionTemplate< T >::train ( const std::vector< trainingExampleTemplate< T > > &  trainingSet )

overridevirtual

Train on a specified set, causes creation if not created

Reimplemented from modelSet< T >.

---

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

• src/regression.h
• src/regression.cpp