Hi Saleem,
When you fit a TH1F histogram (for example named "hist") that has N events with a Gaussian:
hist->Fit("gaus");
the Gaussian function is normalized such that it will give you the number of events inside the histogram.
Doing some math, if G is a Gaussian with some constant term (c), mean (m) and sigma (s) then you have:
G(x) = c * exp ( -0.5 * [(x-m)/s]^2 )
and
Integral(-infinity -> infinity) [G(x)] = c * s * sqrt(2*pi) = N
where: pi = 3.141592654.
In your case Area = N so you get the formula: Area = c * s * sqrt(2*pi)
That's for the built-in function, if you want to get the area directly then create your own Gaussian and define it as:
G(x) = { c / [ s * sqrt(2*pi) ] } * exp ( -0.5 * [(x-m)/s]^2 )
In this case the constant term c is the area directly.
Hope that helps.
Stelios.
----- Original Message -----
From: "Saleem, Muhammad" <saleem@SLAC.stanford.edu>
To: <roottalk@pcroot.cern.ch>
Sent: Wednesday, July 30, 2003 6:51 PM
Subject: [ROOT] built in gaussian
> Hi
>
> The builtin "gaus" function, when i fit gives me :
> constant
> Mean
> and width(sigma)
>
> how do i get the area under the gaussian.
>
> or why this does not give me the area?
>
> am i doing something wrong?
>
> regards
>
> ---saleem
>
>
>
This archive was generated by hypermail 2b29 : Thu Jan 01 2004 - 17:50:14 MET