DataSetTrainer Class Reference

#include <DataSetTrainer.h>

Inheritance diagram for DataSetTrainer:

Collaboration diagram for DataSetTrainer:

Public Member Functions
	DataSetTrainer (Function *function)
	Constructor. More...
virtual	~DataSetTrainer ()
virtual void	train (DataSet *data, int maxEpisodes=10)
virtual void	trainEpisode (DataSet *data)
	Trains one episode over the dataset. More...
virtual bool	stop ()
virtual void	_doTrainEpisode (DataSet *data)=0
	This method should be overriden by subclasses to train a single episode over the dataset. More...
Public Member Functions inherited from Trainer
	Trainer (Function *function)
	Constructor. More...
virtual	~Trainer ()
int	nEpisodes () const
virtual void	init ()

Additional Inherited Members
Public Attributes inherited from Trainer
Function *	_function
	The function this Trainer is optimizing. More...
int	_nEpisodes
	The number of episodes this trainer went through (read-only). More...

Detailed Description

Trains a function over a dataset.

Constructor & Destructor Documentation

DataSetTrainer::DataSetTrainer

(

Function *

function

)

Constructor.

DataSetTrainer::~DataSetTrainer ( )

virtual

Member Function Documentation

virtual void DataSetTrainer::_doTrainEpisode ( DataSet *  data )

pure virtual

This method should be overriden by subclasses to train a single episode over the dataset.

Implemented in BatchRLTrainer, and StochasticDataSetTrainer.

virtual bool DataSetTrainer::stop ( )

inlinevirtual

Should return true iff the training is ended (can be used to provide early-stopping capability). This method is meant to be overriden by subclasses.

void DataSetTrainer::train ( DataSet *  data, int  maxEpisodes = 10  )

virtual

Trains the function over a dataset. Training will stop when the stop() method returns true OR when the number of episodes reaches maxEpisodes. If maxEpisodes <= 0 is specified then it will not be considered (ie. it will only stop when stop() returns true).

void DataSetTrainer::trainEpisode ( DataSet *  data )

virtual

Trains one episode over the dataset.

---

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

• src/qualia/learning/DataSetTrainer.h
• src/qualia/learning/DataSetTrainer.cpp