[R] How to calculate errors in histogram values

[Kilian Hagemann]

Hi there,

I'm new to R but I thought this is the most likely place I could get advice or 
hints w.r.t the following problem:

I have a series of measurements xi with associated uncertainties dxi. I would 
like to construct the probability density histogram of this data where each 
density estimate has an associated error that is derived from the dxi. In 
other words, for large dxi the histogram should also display large 
uncertainties and vice versa. I need this for a curve fitting algorithm.

I have seen many crude ways of working out the error in each bin based on the 
bin count alone, but that's obviously independent of the dxi and thus not 
what I'm after.

So,

1) Is there an R package that can do this (there's nothing in the refence of 
2.1.1)? If so, what algorithm does it use?

2) Could anybody please point me in the right direction (papers, books, 
websites etc.)

Thanks,

-- 
Kilian Hagemann

Climate Systems Analysis Group
University of Cape Town
Republic of South Africa
Tel(w): ++27 21 650 2748