[This is preliminary documentation and is subject to change.]

Using the specified trainer, this method trains the target ITrainableImplicitComponent with the specified settings over the specified datasets

Namespace: Clarion
Assembly: CLARIONLibrary (in CLARIONLibrary.dll) Version: 6.1.0.7 (6.1.0.7)

Syntax

C#
public static double Train<T>(
	T target,
	ImplicitComponent trainer,
	IEnumerable<ActivationCollection> dataSets,
	ImplicitComponentInitializer..::..TrainingTerminationConditions termination,
	int numIterations,
	double sumSqErrThreshold,
	double selection_temperature,
	bool randomTraversal,
	bool testOnly
)
where T : ImplicitComponent, ITrainable

Parameters

target
Type: T
The implicit component that is being trained
trainer
Type: Clarion.Framework.Templates..::..ImplicitComponent
The implicit component that is acting as the trainer (i.e., by providing the DesiredOutput, given each training dataset, to the target)
dataSets
Type: System.Collections.Generic..::..IEnumerable<(Of <(<'ActivationCollection>)>)>
The set of all "input patterns" that are to be used for training the target implicit component
Note
For any input nodes that have an associated ImplicitComponentInitializer..::..Range defined (using AddRange(IWorldObject, Double, Double, Double)), training will occur over that range irrespective of the activations that are specified by the datasets (for those nodes)
termination
Type: Clarion..::..ImplicitComponentInitializer..::..TrainingTerminationConditions
(Optional) The condition under which training should be terminated
numIterations
Type: System..::..Int32
(Optional) If the FIXED termination condition is being used, this specifies the number of times over which the training datasets should be traversed
sumSqErrThreshold
Type: System..::..Double
(Optional) If the SUM_SQ_ERROR termination condition is being used, this specifies the threshold under which the sum of squared error must fall
selection_temperature
Type: System..::..Double
(Optional) If the target implements the IReinforcementTrainable interface, this specifies the temperature for stochastic selection
randomTraversal
Type: System..::..Boolean
(Optional) Indicates whether the data sets should be traversed in a random order for each iteration
testOnly
Type: System..::..Boolean
(Optional) Specifies that the target is ONLY being tested by this call to check its performance based on the specified data sets

Type Parameters

T
The type of ITrainable, ImplicitComponent that is to be trained

Return Value

The accuracy in which the target is performing on the specified data sets

See Also