SG++
python.learner.TrainingStopPolicy.TrainingStopPolicy Class Reference

The class implements heuristics for testing if the learn process should be finished before learner is overfitted. More...

Inheritance diagram for python.learner.TrainingStopPolicy.TrainingStopPolicy:

Public Member Functions

def __init__ (self)
 Contructor. More...
 
def fromJson (cls, jsonObject)
 Restores the TrainingStopPolicy object from the json object with attributes. More...
 
def getAccuracyLimit (self)
 Returns the accuracy on validation data, that have to be achieved. More...
 
def getAdaptiveIterationLimit (self)
 Returns the maximal number of refinement iterations. More...
 
def getEpochsLimit (self)
 Returns the maximal number of iterations, during which accuracy can decreases. More...
 
def getGridSize (self, learner)
 
def getGridSizeLimit (self)
 Returns the maximal grid size. More...
 
def getMSELimit (self)
 Returns MSE on validation data, that have to be achieved. More...
 
def hasGridSizeChanged (self, learner)
 
def hasLimitReached (self, learner)
 
def isTrainingComplete (self, learner)
 Checks if learning process have to be stopped. More...
 
def setAccuracyLimit (self, limit)
 Setter for accuracy limit. More...
 
def setAdaptiveIterationLimit (self, limit)
 
def setEpochsLimit (self, limit)
 Setter for epochs limit. More...
 
def setGridSizeLimit (self, limit)
 Setter for maximal grid size. More...
 
def setMSELimit (self, limit)
 Setter for MSE limit. More...
 
def toString (self)
 Returns a string that represents the object. More...
 

Detailed Description

The class implements heuristics for testing if the learn process should be finished before learner is overfitted.

The test is made by calling method isTrainingComplete(learner) of the class, which returns True if training process should be finished.

Constructor & Destructor Documentation

def python.learner.TrainingStopPolicy.TrainingStopPolicy.__init__ (   self)

Contructor.

References python.learner.TrainingStopPolicy.TrainingStopPolicy.__accuracyLimit, python.learner.TrainingStopPolicy.TrainingStopPolicy.__adaptiveIterationLimit, python.learner.TrainingStopPolicy.TrainingStopPolicy.__epochsLimit, python.learner.TrainingStopPolicy.TrainingStopPolicy.__gridSizeLimit, python.learner.TrainingStopPolicy.TrainingStopPolicy.__MSELimit, and python.learner.TrainingStopPolicy.TrainingStopPolicy.__oldGridSize.

Member Function Documentation

def python.learner.TrainingStopPolicy.TrainingStopPolicy.fromJson (   cls,
  jsonObject 
)

Restores the TrainingStopPolicy object from the json object with attributes.

Parameters
clspython keyword (do not specify)
jsonObjectA json object.
Returns
The restored TrainingStopPolicy object.

Referenced by python.uq.sampler.asgc.ASGCSampler.ASGCSampler.setMemento(), python.uq.analysis.asgc.ASGCKnowledge.ASGCKnowledge.setMemento(), python.uq.learner.Learner.Learner.setMemento(), and python.uq.uq_setting.UQSetting.UQSetting.setMemento().

def python.learner.TrainingStopPolicy.TrainingStopPolicy.getAccuracyLimit (   self)

Returns the accuracy on validation data, that have to be achieved.

Returns
: accuracy on validation data, that have to be achieved

References python.learner.TrainingStopPolicy.TrainingStopPolicy.__accuracyLimit.

def python.learner.TrainingStopPolicy.TrainingStopPolicy.getAdaptiveIterationLimit (   self)

Returns the maximal number of refinement iterations.

Returns
: the maximal number of refinement iterations

References python.learner.TrainingStopPolicy.TrainingStopPolicy.__adaptiveIterationLimit.

def python.learner.TrainingStopPolicy.TrainingStopPolicy.getEpochsLimit (   self)

Returns the maximal number of iterations, during which accuracy can decreases.

Returns
: the maximal number of iterations, during which accuracy can decreases

References python.learner.TrainingStopPolicy.TrainingStopPolicy.__epochsLimit.

def python.learner.TrainingStopPolicy.TrainingStopPolicy.getGridSize (   self,
  learner 
)
def python.learner.TrainingStopPolicy.TrainingStopPolicy.getGridSizeLimit (   self)

Returns the maximal grid size.

Returns
: maximal grid size

References python.learner.TrainingStopPolicy.TrainingStopPolicy.__gridSizeLimit.

Referenced by python.uq.sampler.asgc.ASGCSamplerStopPolicy.ASGCSamplerStopPolicy.hasLimitReached(), and python.learner.TrainingStopPolicy.TrainingStopPolicy.hasLimitReached().

def python.learner.TrainingStopPolicy.TrainingStopPolicy.getMSELimit (   self)

Returns MSE on validation data, that have to be achieved.

Returns
: MSE on validation data, that have to be achieved

References python.learner.TrainingStopPolicy.TrainingStopPolicy.__MSELimit.

Referenced by python.learner.TrainingStopPolicy.TrainingStopPolicy.hasLimitReached().

def python.learner.TrainingStopPolicy.TrainingStopPolicy.hasGridSizeChanged (   self,
  learner 
)
def python.learner.TrainingStopPolicy.TrainingStopPolicy.hasLimitReached (   self,
  learner 
)
def python.learner.TrainingStopPolicy.TrainingStopPolicy.isTrainingComplete (   self,
  learner 
)

Checks if learning process have to be stopped.

Parameters
learnerLearner object
Returns
: boolean value, true if learning has to stop, false otherwise

References python.learner.TrainingStopPolicy.TrainingStopPolicy.__oldGridSize, python.learner.TrainingStopPolicy.TrainingStopPolicy.getGridSize(), python.learner.TrainingStopPolicy.TrainingStopPolicy.hasGridSizeChanged(), and python.learner.TrainingStopPolicy.TrainingStopPolicy.hasLimitReached().

def python.learner.TrainingStopPolicy.TrainingStopPolicy.setAccuracyLimit (   self,
  limit 
)

Setter for accuracy limit.

Parameters
limitdouble accuracy on validation data, that have to be achieved

References python.learner.TrainingStopPolicy.TrainingStopPolicy.__accuracyLimit.

def python.learner.TrainingStopPolicy.TrainingStopPolicy.setAdaptiveIterationLimit (   self,
  limit 
)

References python.learner.TrainingStopPolicy.TrainingStopPolicy.__adaptiveIterationLimit.

def python.learner.TrainingStopPolicy.TrainingStopPolicy.setEpochsLimit (   self,
  limit 
)

Setter for epochs limit.

Parameters
limitinteger Maximal number of iterations, during which accuracy can decreases

References python.learner.TrainingStopPolicy.TrainingStopPolicy.__epochsLimit.

def python.learner.TrainingStopPolicy.TrainingStopPolicy.setGridSizeLimit (   self,
  limit 
)

Setter for maximal grid size.

Parameters
limitinteger maximal grid size

References python.learner.TrainingStopPolicy.TrainingStopPolicy.__gridSizeLimit.

def python.learner.TrainingStopPolicy.TrainingStopPolicy.setMSELimit (   self,
  limit 
)

Setter for MSE limit.

Parameters
limitdouble minimal MSE on validation data, that have to be achieved

References python.learner.TrainingStopPolicy.TrainingStopPolicy.__MSELimit.

def python.learner.TrainingStopPolicy.TrainingStopPolicy.toString (   self)

Returns a string that represents the object.

Returns
A string that represents the object. )

Referenced by python.uq.learner.Learner.Learner.createMemento().


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