C# Class Encog.Neural.Prune.PruneSelective

Prune a neural network selectively. This class allows you to either add or remove neurons from layers of a neural network. You can also randomize or stimulate neurons. No provision is given for removing an entire layer. Removing a layer requires a totally new set of weights between the layers before and after the removed one. This essentially makes any remaining weights useless. At this point you are better off just creating a new network of the desired dimensions.

Mostrar archivo Open project: encog/encog-silverlight-core Class Usage Examples

Public Methods

Method	Description
ChangeNeuronCount ( int layer, int neuronCount ) : void	Change the neuron count for the network. If the count is increased then a zero-weighted neuron is added, which will not affect the output of the neural network. If the neuron count is decreased, then the weakest neuron will be removed. This method cannot be used to remove a bias neuron.
DetermineNeuronSignificance ( int layer, int neuron ) : double	Determine the significance of the neuron. The higher the return value, the more significant the neuron is.
Prune ( int targetLayer, int neuron ) : void	Prune one of the neurons from this layer. Remove all entries in this weight matrix and other layers. This method cannot be used to remove a bias neuron.
PruneSelective ( BasicNetwork network ) : System	Construct an object prune the neural network.
RandomizeNeuron ( double low, double high, int targetLayer, int neuron ) : void
RandomizeNeuron ( int targetLayer, int neuron ) : void	Assign random values to the network. The range will be the min/max of existing neurons.
StimulateNeuron ( double percent, int targetLayer, int neuron ) : void	Stimulate the specified neuron by the specified percent. This is used to randomize the weights and bias values for weak neurons.
StimulateWeakNeurons ( int layer, int count, double percent ) : void	Stimulate weaker neurons on a layer. Find the weakest neurons and then randomize them by the specified percent.

Private Methods

Method	Description
DecreaseNeuronCount ( int layer, int neuronCount ) : void	Internal function to decrease the neuron count of a layer.
FindWeakestNeurons ( int layer, int count ) : int[]	Find the weakest neurons on a layer. Considers both weight and bias.
IncreaseNeuronCount ( int targetLayer, int neuronCount ) : void	Internal function to increase the neuron count. This will add a zero-weight neuron to this layer.
RandomizeNeuron ( int targetLayer, int neuron, bool useRange, double low, double high, bool usePercent, double percent ) : void	Used internally to randomize a neuron. Usually called from randomizeNeuron or stimulateNeuron.
ReindexNetwork ( ) : void	Creat new index values for the network.

Method Details

ChangeNeuronCount() public method

Change the neuron count for the network. If the count is increased then a zero-weighted neuron is added, which will not affect the output of the neural network. If the neuron count is decreased, then the weakest neuron will be removed. This method cannot be used to remove a bias neuron.

public ChangeNeuronCount ( int layer, int neuronCount ) : void
layer	int	The layer to adjust.
neuronCount	int	The new neuron count for this layer.
return	void

DetermineNeuronSignificance() public method

Determine the significance of the neuron. The higher the return value, the more significant the neuron is.

public DetermineNeuronSignificance ( int layer, int neuron ) : double
layer	int	The layer to query.
neuron	int	The neuron to query.
return	double

Prune() public method

Prune one of the neurons from this layer. Remove all entries in this weight matrix and other layers. This method cannot be used to remove a bias neuron.

public Prune ( int targetLayer, int neuron ) : void
targetLayer	int	The neuron to prune. Zero specifies the first neuron.
neuron	int	The neuron to prune.
return	void

PruneSelective() public method

Construct an object prune the neural network.

public PruneSelective ( BasicNetwork network ) : System
network	Encog.Neural.Networks.BasicNetwork	The network to prune.
return	System

RandomizeNeuron() public method

public RandomizeNeuron ( double low, double high, int targetLayer, int neuron ) : void
low	double	The low-end of the range.
high	double	The high-end of the range.
targetLayer	int	The target layer.
neuron	int	The target neuron.
return	void

RandomizeNeuron() public method

Assign random values to the network. The range will be the min/max of existing neurons.

public RandomizeNeuron ( int targetLayer, int neuron ) : void
targetLayer	int	The target layer.
neuron	int	The target neuron.
return	void

StimulateNeuron() public method

Stimulate the specified neuron by the specified percent. This is used to randomize the weights and bias values for weak neurons.

public StimulateNeuron ( double percent, int targetLayer, int neuron ) : void
percent	double	The percent to randomize by.
targetLayer	int	The layer that the neuron is on.
neuron	int	The neuron to randomize.
return	void

StimulateWeakNeurons() public method

Stimulate weaker neurons on a layer. Find the weakest neurons and then randomize them by the specified percent.

public StimulateWeakNeurons ( int layer, int count, double percent ) : void
layer	int	The layer to stimulate.
count	int	The number of weak neurons to stimulate.
percent	double	The percent to stimulate by.
return	void