SG++-Doxygen-Documentation
sgpp::datadriven::RefinementMonitor Class Referenceabstract

Superclass for refinement monitors. More...

#include <RefinementMonitor.hpp>

Inheritance diagram for sgpp::datadriven::RefinementMonitor:
sgpp::datadriven::RefinementMonitorConvergence sgpp::datadriven::RefinementMonitorPeriodic

Public Member Functions

virtual void pushToBuffer (size_t numberInstances, double currentValidError, double currentTrainError)=0
 Stores the current error values in the buffer. More...
 
virtual size_t refinementsNecessary ()=0
 Checks if the model needs to be refined. More...
 
virtual ~RefinementMonitor ()=default
 Destructor. More...
 

Detailed Description

Superclass for refinement monitors.

They track whether a refinement should happen or not.

Constructor & Destructor Documentation

◆ ~RefinementMonitor()

virtual sgpp::datadriven::RefinementMonitor::~RefinementMonitor ( )
virtualdefault

Destructor.

Member Function Documentation

◆ pushToBuffer()

virtual void sgpp::datadriven::RefinementMonitor::pushToBuffer ( size_t  numberInstances,
double  currentValidError,
double  currentTrainError 
)
pure virtual

Stores the current error values in the buffer.

If the buffer has reached the maximum size, the oldest values are removed.

Parameters
numberInstancesthe number of instances added
currentValidErrorThe current validation error
currentTrainErrorThe current training error

Implemented in sgpp::datadriven::RefinementMonitorConvergence, and sgpp::datadriven::RefinementMonitorPeriodic.

Referenced by sgpp::datadriven::RefinementHandler::checkRefinementNecessary(), sgpp::datadriven::SparseGridMinerCrossValidation::learn(), sgpp::datadriven::SparseGridMinerSplitting::learn(), sgpp::datadriven::LearnerSGD::train(), sgpp::datadriven::LearnerSVM::train(), and sgpp::datadriven::LearnerSGDE::trainOnline().

◆ refinementsNecessary()

virtual size_t sgpp::datadriven::RefinementMonitor::refinementsNecessary ( )
pure virtual

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