com.opttek.optquest
Class COptNeuralNetNode

java.lang.Object
  com.opttek.optquest.COptNeuralNetNode

public class COptNeuralNetNode

extends java.lang.Object

The COptNeuralNetNode is used to create nodes for an OptQuest Neural Network. These nodes are added to the COptNeuralNet to create the OptQuest Neural Network. Nodes are defined by a unique node number, a type (input,bias,sum,product), and a level.

Field Summary

static int	BADACTIVATION
static int	BADNODE
static int	BIASNODE
static int	DEFAULTACTIVATION
static int	IDENTITYACTIVATION
static int	INPUTNODE
static int	PRODUCTNODE
static int	SIGMOID01ACTIVATION
static int	SIGMOIDNEG11ACTIVATION
static java.lang.String	STRING_ACTIVATION01
static java.lang.String	STRING_ACTIVATIONDEFAULT
static java.lang.String	STRING_ACTIVATIONIDENTITY
static java.lang.String	STRING_ACTIVATIONMINUS11
static java.lang.String	STRING_TYPEBIAS
static java.lang.String	STRING_TYPEINPUT
static java.lang.String	STRING_TYPEPRODUCT
static java.lang.String	STRING_TYPESUM
static int	SUMNODE

Constructor Summary

COptNeuralNetNode()
Default Constructor.

COptNeuralNetNode(int type, int num, int level)
Constructor that accepts the type, node number and level as input.

Method Summary

int	GetActivationFunction() Get the activation function of the node.
int	GetLevel() Get the level of the node.
double	GetMaximumScale() Get the maximum value of the scaling.
double	GetMinimumScale() Get the minimum value of the scaling.
int	GetNodeNumber() Get the node number of the node.
int	GetType() Get the type of the node.
boolean	IsScaled() Check if the input/output node is scaled.
void	SetActivationFunction(int act) Set the activation function of the node.
void	SetLevel(int level) Set the level of the node.
void	SetNodeNumber(int number) Set the node number of the node.
void	SetScale(double min, double max) Set the scaling of the node.
void	SetType(int type) Set the type of the node.

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

INPUTNODE

public static final int INPUTNODE

See Also:

Constant Field Values

BIASNODE

public static final int BIASNODE

See Also:

Constant Field Values

SUMNODE

public static final int SUMNODE

See Also:

Constant Field Values

PRODUCTNODE

public static final int PRODUCTNODE

See Also:

Constant Field Values

BADNODE

public static final int BADNODE

See Also:

Constant Field Values

STRING_TYPEINPUT

public static final java.lang.String STRING_TYPEINPUT

See Also:

Constant Field Values

STRING_TYPEBIAS

public static final java.lang.String STRING_TYPEBIAS

See Also:

Constant Field Values

STRING_TYPESUM

public static final java.lang.String STRING_TYPESUM

See Also:

Constant Field Values

STRING_TYPEPRODUCT

public static final java.lang.String STRING_TYPEPRODUCT

See Also:

Constant Field Values

DEFAULTACTIVATION

public static final int DEFAULTACTIVATION

See Also:

Constant Field Values

SIGMOID01ACTIVATION

public static final int SIGMOID01ACTIVATION

See Also:

Constant Field Values

SIGMOIDNEG11ACTIVATION

public static final int SIGMOIDNEG11ACTIVATION

See Also:

Constant Field Values

IDENTITYACTIVATION

public static final int IDENTITYACTIVATION

See Also:

Constant Field Values

BADACTIVATION

public static final int BADACTIVATION

See Also:

Constant Field Values

STRING_ACTIVATION01

public static final java.lang.String STRING_ACTIVATION01

See Also:

Constant Field Values

STRING_ACTIVATIONMINUS11

public static final java.lang.String STRING_ACTIVATIONMINUS11

See Also:

Constant Field Values

STRING_ACTIVATIONIDENTITY

public static final java.lang.String STRING_ACTIVATIONIDENTITY

See Also:

Constant Field Values

STRING_ACTIVATIONDEFAULT

public static final java.lang.String STRING_ACTIVATIONDEFAULT

See Also:

Constant Field Values

Constructor Detail

COptNeuralNetNode

public COptNeuralNetNode()

Default Constructor. The type, number, and level must be set before this is a valid node.

COptNeuralNetNode

public COptNeuralNetNode(int type,
                         int num,
                         int level)

Constructor that accepts the type, node number and level as input. Each node number must be unique within the network. The level is used to determine how the network is traversed

Parameters:

type - - the type of the node.

num - - the number of the node.

level - - the level of the node.

Method Detail

SetNodeNumber

public void SetNodeNumber(int number)

Set the node number of the node. Each node number must be unique within the network.

Parameters:

number - - the number of the node.

GetNodeNumber

public int GetNodeNumber()

Get the node number of the node.

Returns:

node number of the node

SetLevel

public void SetLevel(int level)

Set the level of the node. The level is used to determine how the network is traversed; the network is traversed from level 1 to the last level in the network one level at a time.

Parameters:

level - - the level of the node.

GetLevel

public int GetLevel()

Get the level of the node.

Returns:

level of the node

SetType

public void SetType(int type)

Set the type of the node.

Parameters:

type - - the type of the node.

• COptNeuralNetNode.INPUTNODE. The node is an input node whose value is provided by the input matrix.
• COptNeuralNetNode.BIASNODE. The node is a bias node whose value is 1.0.
• COptNeuralNetNode.SUMNODE. The node is a node whose value is the sum of the incoming arcs.
• COptNeuralNetNode.PRODUCTNODE. The node is a node whose value is the product of the incoming arcs.

GetType

public int GetType()

Get the type of the node.

Returns:

Type of the node

SetActivationFunction

public void SetActivationFunction(int act)

Set the activation function of the node.

Parameters:

act - - the activation function.

• COptNeuralNetNode.DEFAULTACTIVATION. The node uses the default activation function for the entire network. This is the value that is usually used.
• COptNeuralNetNode.SIGMOID01ACTIVATION. The node uses the [0,1] sigmoid function as the activation function.
• COptNeuralNetNode.SIGMOIDNEG11ACTIVATION. The node uses the [-1,1] sigmoid function as the activation function.
• COptNeuralNetNode.IDENTITYACTIVATION. The node uses the identity function as the activation function. No function is applied to the value.

GetActivationFunction

public int GetActivationFunction()

Get the activation function of the node.

Returns:

activation function of the node

SetScale

public void SetScale(double min,
                     double max)

Set the scaling of the node. The scaling is used to scale input or output nodes to make them either a [0,1] or a [-1,1] value.

Parameters:

min - - the minimum value for the input/output.

max - - the maximum value for the input/output.

IsScaled

public boolean IsScaled()

Check if the input/output node is scaled.

Returns:

true if scaled, false if not

GetMinimumScale

public double GetMinimumScale()

Get the minimum value of the scaling.

Returns:

minimum value of the scaling

GetMaximumScale

public double GetMaximumScale()

Get the maximum value of the scaling.

Returns:

maximum value of the scaling