class TMVA::TActivationTanh: public TMVA::TActivation


  Tanh activation function for ANN. This really simple implementation
  uses TFormulas and should probably be replaced with something more
  efficient later.

Function Members (Methods)

public:

virtual	~TActivationTanh()
static TClass*	Class()
virtual Double_t	Eval(Double_t arg)
virtual Double_t	EvalDerivative(Double_t arg)
virtual TString	GetExpression()
virtual Double_t	GetMax()
virtual Double_t	GetMin()
virtual TClass*	IsA() const
virtual void	MakeFunction(ostream& fout, const TString& fncName)
TMVA::TActivationTanh&	operator=(const TMVA::TActivationTanh&)
virtual void	ShowMembers(TMemberInspector& insp)
virtual void	Streamer(TBuffer& b)
void	StreamerNVirtual(TBuffer& b)
TMVA::TActivationTanh	TActivationTanh()
TMVA::TActivationTanh	TActivationTanh(const TMVA::TActivationTanh&)

Data Members

private:

TFormula*	fEqn	equation of tanh sigmoid
TFormula*	fEqnDerivative	equation of tanh sigmoid derivative

Class Charts

Inheritance Inherited Members Includes Libraries

Function documentation

TActivationTanh()

 constructor for tanh sigmoid (normalized in [-1,1])

~TActivationTanh()

 destructor

Double_t Eval(Double_t arg)

 evaluate the tanh

Double_t EvalDerivative(Double_t arg)

 evaluate the derivative

TString GetExpression()

 get expressions for the tanh and its derivative

void MakeFunction(ostream& fout, const TString& fncName)

 writes the sigmoid activation function source code

TActivationTanh()

Double_t GetMin()

 minimum of the range of the activation function

{ return -1; }

Double_t GetMax()

 maximum of the range of the activation function

{ return 1; }