FakeSigmoidFunction Class Reference

Fake Sigmoid Function !! Is a linear approximation of sigmoid function. More...

Inheritance diagram for FakeSigmoidFunction:

List of all members.

Public Member Functions
	FakeSigmoidFunction (double l=1.0)
	Construct a sigmoid updater with parameter l.
virtual	~FakeSigmoidFunction ()
	Destructor.
virtual void	apply (DoubleVector &inputs, DoubleVector &outputs)
	Implement the updating method.
virtual void	configure (ConfigurationParameters &params, QString prefix)
	Configures the object using a ConfigurationParameters object.
virtual bool	derivate (const DoubleVector &x, const DoubleVector &y, DoubleVector &d) const
	return the approximation commonly used in backpropagation learning: x(1-x)
virtual void	save (ConfigurationParameters &params, QString prefix)
	Save the actual status of parameters into the ConfigurationParameters object passed.
Public Attributes
double	lambda
	lambda is the slope of the curve

---

Detailed Description

Fake Sigmoid Function !! Is a linear approximation of sigmoid function.

Definition at line 183 of file liboutputfunctions.h.

---

Constructor & Destructor Documentation

FakeSigmoidFunction

(

double

l = 1.0

)

Construct a sigmoid updater with parameter l.

Definition at line 158 of file liboutputfunctions.cpp.

References FakeSigmoidFunction::lambda.

virtual ~FakeSigmoidFunction

(

)

[inline, virtual]

Destructor.

Definition at line 188 of file liboutputfunctions.h.

---

Member Function Documentation

void apply	(	DoubleVector &	inputs,
		DoubleVector &	outputs
	)		[virtual]

Implement the updating method.

Implements OutputFunction.

Definition at line 163 of file liboutputfunctions.cpp.

References FakeSigmoidFunction::lambda, and DoubleVector::size().

void configure	(	ConfigurationParameters &	params,
		QString	prefix
	)		[virtual]

Configures the object using a ConfigurationParameters object.

Parameters:

params	the configuration parameters object with parameters to use
prefix	the prefix to use to access the object configuration parameters. This is guaranteed to end with the separator character when called by the factory, so you don't need to add one

Implements ParameterSettableWithConfigureFunction.

Definition at line 192 of file liboutputfunctions.cpp.

References ConfigurationParameters::getValue(), and FakeSigmoidFunction::lambda.

bool derivate	(	const DoubleVector &	x,
		const DoubleVector &	y,
		DoubleVector &	d
	)		const [virtual]

return the approximation commonly used in backpropagation learning: x(1-x)

Reimplemented from OutputFunction.

Definition at line 184 of file liboutputfunctions.cpp.

References FakeSigmoidFunction::lambda, and farsa::subtract().

void save	(	ConfigurationParameters &	params,
		QString	prefix
	)		[virtual]

Save the actual status of parameters into the ConfigurationParameters object passed.

Parameters:

params	the configuration parameters object on which save actual parameters
prefix	the prefix to use to access the object configuration parameters.

Implements ParameterSettable.

Definition at line 205 of file liboutputfunctions.cpp.

References ConfigurationParameters::createParameter(), FakeSigmoidFunction::lambda, and ConfigurationParameters::startObjectParameters().

---

Member Data Documentation

double lambda

lambda is the slope of the curve

Definition at line 213 of file liboutputfunctions.h.

Referenced by FakeSigmoidFunction::apply(), FakeSigmoidFunction::configure(), FakeSigmoidFunction::derivate(), FakeSigmoidFunction::FakeSigmoidFunction(), and FakeSigmoidFunction::save().

---

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

• nnfw/include/liboutputfunctions.h
• nnfw/src/liboutputfunctions.cpp