Suggestion for small addition, TH1::Normalize()

[dvotaw]

It seems like this is a common enough operation that it could be implemented as a function. The simplest version of what I mean would just be:

TH1::Normalize()
{
  this->Scale(1.0/this->Integral());
}

It could be modified to take arguments, throw exceptions (for example if “this” has zero integral), but you get what I mean.

[bellenot]

Thanks for the suggestion. I’ll let @moneta comment on it. BTW, did you see Normalizing histograms in the TH1 Class Reference?

[moneta]

I agree this could be an useful function, although we have the TH1::Scale function. In that case it would be useful also the option to divide by 1.0/ (Integral * binWidth) in case the histogram is interpreted as a probability density function.
If you want you can create a Pull Request for this in GitHub

Lorenzo

[dvotaw]

I’ve made a pull request (hopefully I’ve done it correctly).

I just scaled by 1/Integral rather than 1/(Integral * binWidth), because that’s what I’ve used more often in practice.

Maybe it could be modified to take an Option_t argument and do one or the other. I just wasn’t sure how to implement it for 1/(Integral * binWidth), in the case where different bins have different widths.

[Wile_E_Coyote]

Note that there exist at least 7 ways how to do it so, if you really want to implement a new method, it should cover all of them:

Different ways of normalizing histograms

Hi everyone, I’m a new user of both ROOT and RootTalk so please pardon my ignorance. I’ve been trying to normalize several histograms, and when I search in RootTalk, a multitude of ways comes up. How do I know which one is “right”? What is the difference between the following methods? Here are the 4 methods I tried: Method 1: Double_t num = h->GetBinContent(i); Double_t den = h->GetBinWidth(i); Double_t value = 0; if (den!=0) { value = num/den; h->SetBinContent(i,value); } Met…

[moneta]

Hi ,
Many of the different ways mentioned in that posts are not correct, e.g. using histogram entries, only SetBinContent and not SetBinError, etc…
I think for this reason is worth having a function which calls TH1::Scale in the correct way.
There are only 2 cases:

• normalise by the total counts (integral) to show the frequency probability in each bin
• normalise by the total counts * bin width to show the estimated probability density function

If we want to be more precise there is also the question of underflow/overflow.
When normalizing by the total counts we could in principle include underflow/overflow, while we cannot do in the second case because we don’t know the underflow/overflow bin width.

Lorenzo