TDecompSVD not working

Roofit and RooStats
1

Hi all,

I am trying to use TDecompSVD to solve Ax=b, where I have A saved as a TH2 with 300x300 bins. And I have b saved as a TH1 with 300 bins.

{

  TH2D myhist("h1","2DGaussian",300,-3,3,300,-3,3);
 TRandom3 foo(12345);
 for(int i=0;i<10000;i++)
     {
      myhist->Fill(foo.Gaus(0,1),foo.Gaus(0,1));
     }

  const int exbins=myhist->GetNbinsX();
  const int whybins=myhist->GetNbinsY();

  TMatrixD Amatrix(exbins,whybins);
  Amatrix.Zero();
  for (int xbin=1;xbin<=exbins;xbin++)
    {
      for (int ybin=1;ybin<=whybins;ybin++)
	{
          Amatrix(whybins-ybin,xbin-1)=myhist->GetBinContent(xbin,ybin);
        }//end of loop over y bins                                                                                            
    }//end of loop over xbins                                                                                                 

  TDecompSVD svd(Amatrix,0.00001); //Decomp my A matrix with a tolerance of 1e-5
  
  Bool_t ok;
  
  TH1D *hdata = new TH1D("mydataexample","mydataexample",300,0.01,10);
  for (int i=0;i<10000;i++)
     {
        hdata->Fill(foo.Gaus(2,5));
     }

    const int exdata=hdata->GetNbinsX();  
    
    TMatrixD matdata(exdata,1);
    matdata.Zero();
    
  for (int xbin=1;xbin<=exdata;xbin++)
    {
      matdata(xbin-1,0)=hdata->GetBinContent(xbin);
    }//end of loop over xbins  


   const TVectorD c_svd = svd.Solve(matdata,ok);
   
   }

This throws an error, does anyone have any ideas?

2

Hi,

Your code is not correrct. Attached is after the correction and it is working for me

Best

Lorenzo

Untitled.C (1.3 KB)