Implementing rolling average of data (using a normalised gaussian function)

[Mansi]

Hello,

I am trying to implement in my code to have a routine for calculating the rolling average or moving weighted average using a normalised gaussian function.

The idea is as follows -

The weights are given by the chosen distribution, in this case Gaussian.
The weight would be given by evaluating a gaussian for the current bin: i.e. the weight is proportional to exp(-(x-)^2/2*sigma^2).
At each bin, I perform a loop over all neighboring bins.
The centroid of the current bin would be “”.
The “x” would be the centroid of the neighboring bin within the loop.

in the code I try it like this-

  int nbinsIp = hIntensity2D->GetXaxis()->GetNbins();
		  Double_t ReEvalIp;
		  Double_t Ygaus;
		   for(int i =0;nbinsIp+1;i++)
		    {
		      // //re-set the re-evaluated y-value for the current bin
		       ReEvalIp=0;
		       double bcx = hIntensity2D->GetXaxis()->GetBinCenter(i);
		       funcGaus->SetParameter(1,bcx);// set mean to bin center , centroid of the current bin
		       //loop over all neighboring bins
		       for (int j =0; j<i; j++)
		       	{ //evaluate the gaussian function at the neighboring bin
		       	  Ygaus = funcGaus->Eval(hIntensity2D->GetXaxis()->GetBinCenter(j));
		      	  ReEvalIp = +Ygaus;
			  

			}
		  }
  

I am not sure if I am doing it correctly. The code gets stuck forever in this loop.
Can anyone help out with the rolling average method or its implementation.

Thanks

---

Please read tips for efficient and successful posting and posting code

ROOT Version: 6.18
Platform: Mac OS Big Sur
Compiler:

---

[Mansi]

hello,

if anyone has implemented the rolling average method, please do reach out.

thanks

[joa]

Hi,

The conditions to continue the first for loop is always true. Missing a “i<” I guess.

Joa

[eliotoscar]

I am trying to implement in my code to have a routine for calculating the rolling average or moving weighted average using a normalised gaussian function.

The weights are given by the chosen distribution, in this case Gaussian.
The weight would be given by evaluating a gaussian for the current bin: i.e.