[R] hist densities

[Prof Brian Ripley]

On Fri, 21 Mar 2008, Luke Spadavecchia wrote:

> Hi there,
>
> I was wondering why
>
> x<-rnorm(n=5000,mean=1.5,sd=0.25)
> h<-hist(x)
> sum(h$density)
>
> [1] 10
>
> I thought the integral of the histogram should be one? These seems
> strange to me, or am I being silly...

The integral is not a sum of heights but of areas.  You need
sum(h$density * diff(h$breaks)), and this is spelled out on the help page.

> Regards Luke Spadavecchia
>
> PS. I'm using R 2.5.1 on Mac OSX

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595