In processing astronomical images (e.g., Hubble), there will be cosmic ray
hits (a single or cluster of pixels) with very high counts. So to set the
background level, the median will not be strongly affected by a few such
pixels, but the mean would be rendered useless. Just one example of
where the the median is quite useful.
Patrick
On Mon, 19 Oct 2009, Sherlock, Ric wrote:
Yes, for example the median is a better measurement of central location than
the mean for skewed distributions (incomes, test results).
From: [email protected]
Tukey did show that median and quantiles are often much more stable
than average.( f.e median polish and median smoothing in general)
-------------------
From: Bo Jacoby
We are solving a problem which should not be solved. The advantage
of Median and Quantiles, as opposed to Arithmetical Mean and Standard
Deviation, is that they can be evaluated without computing. Having a
computer at hand, this argument vanishes.
--- Den s?n 18/10/09 skrev Fraser Jackson
Statistical programs have a range of
alternatives for the quantile function.
The following script embodies forms
considered in a useful survey paper some years ago. It
requires further
functions to answer Devon's question but
does include some options worth considering.
NB. Quantile functions
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
