Fit on one data with one implicit function

Version Info

Minimum Origin Version Required: OriginPro 9.0 SR0

Examples

The following code shows how to use NLFit to fit with ellipse data.

1. Create a new worksheet and import the data <Origin Installation Directory>\Samples\Curve Fitting\Ellipse.dat.
2. Copy and compile the following codes in Code Builder, and then run "NLFit_Implicit_example1" in Command Window.

//////////////////////////////////////////////////////////////////////////////////////
// Implicit function fitting, fitting a ellipse f(x,y) = 1 - ((x - xc)/a)^2 - ((y-yc)/b)^2
//////////////////////////////////////////////////////////////////////////////////////
#include <Origin.h>
#include <FDFTree.h>
#include <ONLSF.h>

void       NLFit_Implicit_example1()
{
        string                  strFDF =  okutil_get_origin_path(ORIGIN_PATH_SYSTEM, "FitFunc") + "ellipse.FDF"; 
        Tree                    trFF;
        if( !nlsf_FDF_to_tree( strFDF, &trFF ))
        {
                out_str("Fail to load FDF function file");
                return;
        }
        
        ///////////////////////////////////////////////////////////////////////////
        // 1. Setting the fitting function
        
        // The fit object
        NLFit                   fit;
        int                                nn = fit.SetFunction(trFF);
        if (nn <= 0)
        {
                out_str("Failed setting fitting function!");
                return;
        }
  
        ///////////////////////////////////////////////////////////////////////////
        // 2. Setting data for fitting ////////////////////////////////////////////
        Worksheet        wks = Project.ActiveLayer();
        if( !wks )
        {
                out_str("No active worksheet to get input data");
                return;
        }
        
        vector vX = wks.Columns(0).GetDataObject();
        vector vY = wks.Columns(1).GetDataObject();
        vector vWx = wks.Columns(2).GetDataObject();
        vector vWy = wks.Columns(3).GetDataObject();
        
        NLSFONEDATA     stIndep[2];
        stIndep[0].pdData = vX;
        stIndep[1].pdData = vY;        
        stIndep[0].nSize = stIndep[1].nSize = vX.GetSize();
        
        ONEDATAWEIGHT   stWeightIndep[2];
        stWeightIndep[0].stWeightData.pdData  = vWx;
        stWeightIndep[0].stWeightData.nSize = vWx.GetSize();
        stWeightIndep[0].method = WEIGHT_DIRECT;
        stWeightIndep[0].a = stWeightIndep[0].b = stWeightIndep[0].c = 0;
        
        stWeightIndep[1].stWeightData.pdData  = vWy;
        stWeightIndep[1].stWeightData.nSize = vWy.GetSize();
        stWeightIndep[1].method = WEIGHT_DIRECT;
        stWeightIndep[1].a = stWeightIndep[1].b = stWeightIndep[1].c = 0;
                
        nn = fit.SetData(1, NULL, &stIndep, NULL, &stWeightIndep);
        if(nn != 0)
        {
                out_str("SetData() failed!");
                return;
        }
        
        // 3. Setting paramaters //////////////////////////////////////////////////
        int                                nNumParams = 4;       
        vector                  vParams(nNumParams);  // this vector should be initialized to the total number of paramaters
        // This vector will be used both to set initial paramater values, and to receive the parameter values after fitting:
        vParams[0] = 1;                //xc
        vParams[1] = 2;                //yc
        vParams[2] = 3;                //a
        vParams[3] = 2;                //b
        nn = fit.SetParams(nNumParams, vParams);
        if ( 0 != nn )
        {
                out_str("Invalid paramater setting!");
                return;
        }
 
 
        ///////////////////////////////////////////////////////////////////////////
        // 4. Fitting /////////////////////////////////////////////////////////////
        int                                nMaxNumIterations = 100;
        nn = fit.Fit(nMaxNumIterations, NULL, NULL, NULL, FITMETH_ORTHOGONAL_DISTANCE_REGRESSION);
        printf("Fit(%ld) returned: %ld\n", nMaxNumIterations, nn);
        if( 0 == nn )
                printf("Fit Done!\n");    
 
        /////////////////////////////////////////////////////////////////////////////
        // 5. Dump all the results and compare with what was expected /////////////
        for (int iparam = 0; iparam < nNumParams; iparam++)
        {
                printf("param index = %d   value obtained = %lf\n", iparam + 1, vParams[iparam]);
        }
 
        return;    
}