virtual double | DoDerivative(double x) const |
virtual double | DoEval(double x) const |
virtual double | DoEvalPar(double x, const double* p) const |
virtual double | DoParameterDerivative(double x, const double* p, unsigned int ipar) const |
TF1* | fFunc | pointer to ROOT function |
bool | fLinear | flag for linear functions |
vector<double> | fParams | cached vector with parameter values |
bool | fPolynomial | flag for polynomial functions |
double | fX[1] | ! cached vector for x value (needed for TF1::EvalPar signature) |
static double | fgEps | epsilon used in derivative calculation h ~ eps |x| |
evaluate the derivative of the function with respect to the parameters
evaluate the derivative of the function with respect to the parameters IMPORTANT NOTE: TF1::GradientPar returns 0 for fixed parameters to avoid computing useless derivatives BUT the TLinearFitter wants to have the derivatives also for fixed parameters. so in case of fLinear (or fPolynomial) a non-zero value will be returned for fixed parameters
@name interface inherited from IFunction Clone the wrapper but not the original function
@name interface inherited from IParamFunction get the parameter values (return values cachen inside, those inside TF1 might be different)
set parameter values (only the cached one in this class,leave unchanges those of TF1)
calculate function and derivative at same time (required by IGradient interface)
evaluate function passing coordinates x and vector of parameters
evaluate function using the cached parameter values of this class (not of TF1) re-implement for better efficiency