The SDL Component Suite is an industry leading collection of components supporting scientific and engineering computing. Please visit the SDL Web site for more information....


Home MathPack Math2 Class TCurveFit Properties ModelPar
See also: CalcRegModel, CenteredPolynomial, PolyOrder, CPXShift 

ModelPar

Unit: SDL_math2
Class: TCurveFit
Declaration: property ModelPar [ix,iy: integer]: double;

The read-only two-dimensional array property ModelPar returns the model parameters of the previously calculated regression model. Depending on the type of regression model the ModelPar array contains different parameters.

The parameter ix may take the values 1 and 2: for ix=1 the actual model parameter, for ix=2 the corresponding p value of this parameter is returned. Please note that the p values are only calculated for polynomials and centered polynomials, for all other models the p values are set to -1 in order to indicate that they are not available.

Following is the list of parameters of the various regression models:

RegModel Parameter Regression Function Details Availability of p Values
rmLinear k0... ModelPar[1,0]
k1... ModelPar[1,1] 		no (alternative: polynomial fit with order 1)
rmParabolic k0... ModelPar[1,0]
k1... ModelPar[1,1]
k2... ModelPar[1,2] 		no (alternative: polynomial fit with order 2)
rmGaussian k0... ModelPar[1,0]
k1... ModelPar[1,1]
k2... ModelPar[1,2] 		no
rmReciLin k0... ModelPar[1,0]
k1... ModelPar[1,1] 		no
rmHyperbolic k0... ModelPar[1,0]
k1... ModelPar[1,1] 		no
rmReciHyperbolic k0... ModelPar[1,0]
k1... ModelPar[1,1] 		no
rmLog k0... ModelPar[1,0]
k1... ModelPar[1,1] 		no
rmReciLog k0... ModelPar[1,0]
k1... ModelPar[1,1] 		no
rmPower k0... ModelPar[1,0]
k1... ModelPar[1,1] 		no
rmExpo k0... ModelPar[1,0]
k1... ModelPar[1,1] 		no
rmHoerl k0... ModelPar[1,0]
k1... ModelPar[1,1] 		no
rmCircle dx ... ModelPar[1,0]
dy ... ModelPar[1,1]
r ... ModelPar[1,2] 		no
rmPolynomial k0... ModelPar[1,0]
k1... ModelPar[1,1]
...
kn... ModelPar[1,n]
n ... PolyOrder 		yes:
p0 = ModelPar[2,0]
...
pn = ModelPar[2,n]
rmCenteredPoly k0... ModelPar[1,0]
k1... ModelPar[1,1]
...
kn... ModelPar[1,n]
d ... CPXShift
n ... PolyOrder 		yes:
p0 = ModelPar[2,0]
...
pn = ModelPar[2,n]

Last Update: 2023-Feb-06