TMtxMultiNonLinReg Class

class TMtxMultiNonLinReg : public TMtxComponent;

Performs multiple-nonlinear regression.

How to use TMtxMultiNonLinReg component?
1. Drop a TMtxMultiNonLinReg component on the form. 2. Define TMtxMultiNonLinReg::X the vector of independent variables. 3. Define TMtxMultiNonLinReg::Y the vector of dependent variables. Make sure that X and Y have the same length. 4. Define initial estimates for TMtxMultiNonLinReg::B regression parameters. 5. (Optionaly) define TMtxMultiNonLinReg::Weights. If you'll use weights, set TMtxMultiNonLinReg::UseWeights property to true. 6. Define regression function. 7. (Optionally) define the derivative procedure. If you don't define derivative procedure then the numeric approximation will be used to calculate the derivative at specific point. 8. Define the optimization method. Depending on your function different methods will give better/worse results. 9. (Optionally) define desired tolerance for result. 10. Define the maximum number of steps for optimization method. Internal optimization method will stop when number of iterations exceeds MaxIter or internal minimum precision is within Tolerance. 11. Verbose property, if assigned, stores Fun, evaluated at each iteration step. Optionally, you can also assign TOptControl object to the Verbose property. This allows the optimization procedure to be interrupted from another thread and optionally also allows logging and iteration count monitoring.

Results:
1. TMtxMultiNonLinReg::B : Regression coefficients estimates. 2. TMtxMultiNonLinReg::YCalc : Fitted Y values.

The component internally performs vectorized optimization process. It can also be used to speed-up nonlinear regression with only one independent variable by 10-20x.

Examples

cpp

#include "MtxVec.h"
#include "Optimization.h"
#include "Regress.h"
#include "StatTools.h"

using namespace Dew;
using namespace Dew::Math;
using namespace Dew::Stats;

static void Rat43(TVec* b, TVecList* x, TVec* y)
{
//        return b[0] / Math.Pow((1.0 + Math.Exp(b[1]-b[2]*x)),1/b[3]);

    y->Mul((*x)[0], -(*b)[2]);
    y->Add((*b)[1]);
    y->Exp();
    y->Add(1);
    y->Power(1/(*b)[3]);
    y->DivideBy((*b)[0]);
}

void Example()
{
    TMtxMultiNonLinReg* nlr = new TMtxMultiNonLinReg(nullptr);
    // Load data - independent variable
    nlr->X->Add();
    (*nlr->X)[0]->SetIt(false,DewArray<double> {9.000, 14.000, 21.000, 28.000,
        42.000, 57.000, 63.000, 70.000,
        79.000});
    // Load data - dependent variable
    nlr->Y->SetIt(false,DewArray<double> {8.930, 10.800, 18.590, 22.330,
        39.350, 56.110, 61.730, 64.620,
        67.080});
    // Initial estimates for regression coefficients
    nlr->B->SetIt(false,DewArray<double> {100,10,1,1});
    // setup optimization parameters
    nlr->Tolerance = 1.0e-6; // 6 digits should do the trick
    nlr->GradTolerance = 1.0e-3; // 3 digits
    nlr->MaxIteration = 400;
    nlr->RegressFunction = Rat43; // regression function
    // Marquardt method
    nlr->OptMethod = TOptMethod::optMarquardt;
    nlr->Recalc();
    // MtxNinLinReg->b now stores calculated regression parameter estimates
    delete nlr;
}

Constructors

Name	Description
TMtxMultiNonLinReg (2)

Destructors

Name	Description
~TMtxMultiNonLinReg

Properties

Name	Type	Description
AutoUpdate	bool	If true then changing any of the TMtxMultiNonLinReg properties will trigger the TMtxMultiNonLinReg::Recalc method.
B	TVec*	Stores the regression coefficients.
BlockAssign	bool	Block streaming of specific properties when storing only a "template".
DeriveProcedure	TMultiDeriveProc	Defines derivatives of regression function.
Dirty	bool	When any of the TMtxNonLinReg properties changes, this property is automatically set to true.
EditorActive	bool	Returns True, if the component editor is displayed.
GradTolerance	double	Defines minimal allowed gradient C-Norm.
MaxIteration	int	Defines one of the conditions for terminating the regression coefficient calculation.
Name	DewString
OptMethod	TOptMethod	Optimization method used.
Reference	TReferenceList*	Stores a list of components that have to be notified, when this component is destroyed.
RegressFunction	TMultiRegressFun	Defines the regression function of several regression parameters.
SoftSearch	bool	Defines line search algorithm for Quasi-Newton and Conjugate methods.
StopReason	TOptStopReason	Returns the reason why regression parameters calculation stopped.
Tolerance	double	Defines one of the conditions for terminating the regression coefficient calculation.
UseWeights	bool	Weighted non-linear regression.
Verbose	tStrings*	If not nil then the optimization method uses it for logging each optimization step.
Weights	TVec*	Weights.
X	TVecList*	Vector of independant variables.
Y	TVec*	Vector of dependant variables.
YCalc	TVec*	Vector of calculated values.

Methods

Name	Description
AllowStreaming	Returns the negatated value of BlockAssign.
Assign	Assign values of all published properties from Source.
AssignTemplate	Assign values of "template-like" published properties from Source.
EditorClass	To be overriden in descendanr classses.
LoadFromStream	Load the component from Src stream.
LoadTemplateFromFile	Load a template from file.
LoadTemplateFromStream	Load a template from stream.
Recalc	Triggers non-linear regression recalculation.
ReferenceRemoved
Reset	Reset is called after loading the data from stream or file.
SaveTemplateToFile	Save template to file.
SaveTemplateToStream	Save template to stream.
SaveToStream	Save the component to Dst stream.

Fields

Name	Type	Description
Iterations	int	Number of iterations needed to converge to solution.
Tag	intptr_t