A monitor that decides whether refinements should be performed using a simple periodic approach: After at least a certain amount of instances has arrived, the monitor will allow a new refinement. More...

#include <RefinementMonitorPeriodic.hpp>

Inheritance diagram for sgpp::datadriven::RefinementMonitorPeriodic:

Public Member Functions
void	pushToBuffer (size_t numberInstances, double currentValidError, double currentTrainError) override
	Pushes a new iteration to the monitors state. More...

	RefinementMonitorPeriodic (size_t period)
	Constructor for the periodic refinement monitor. More...

size_t	refinementsNecessary () override
	Checks if the model needs to be refined. More...

virtual	~RefinementMonitorPeriodic ()
	Destructor. More...

Public Member Functions inherited from sgpp::datadriven::RefinementMonitor
virtual	~RefinementMonitor ()=default
	Destructor. More...

Detailed Description

A monitor that decides whether refinements should be performed using a simple periodic approach: After at least a certain amount of instances has arrived, the monitor will allow a new refinement.

Constructor & Destructor Documentation

◆ RefinementMonitorPeriodic()

sgpp::datadriven::RefinementMonitorPeriodic::RefinementMonitorPeriodic ( size_t period )

explicit

Constructor for the periodic refinement monitor.

Parameters

period the number of instances that is needed to trigger a new refinement. If set to zero, this monitor will always trigger exactly one refinement

◆ ~RefinementMonitorPeriodic()

sgpp::datadriven::RefinementMonitorPeriodic::~RefinementMonitorPeriodic ( )

virtual

Destructor.

Member Function Documentation

◆ pushToBuffer()

void sgpp::datadriven::RefinementMonitorPeriodic::pushToBuffer	(	size_t	numberInstances,
		double	currentValidError,
		double	currentTrainError
	)

overridevirtual

Pushes a new iteration to the monitors state.

Parameters

numberInstances	the number of instances that were used for the training step
currentValidError	The current validation error
currentTrainError	The current training error

Implements sgpp::datadriven::RefinementMonitor.

◆ refinementsNecessary()

size_t sgpp::datadriven::RefinementMonitorPeriodic::refinementsNecessary ( )

overridevirtual

Checks if the model needs to be refined.

If multiples of the period data instances have arrived this monitor might trigger more than one refinement

Returns: the number of refinements that are triggered by the monitor