And that's not to mention those old chestnuts, the mean (location) and
variance (scale)
-----Original Message-----
From: A UK-based worldwide e-mail broadcast system mailing list
[mailto:[log in to unmask]] On Behalf Of Andre Francis
Sent: 24 November 2010 19:15
To: [log in to unmask]
Subject: Re: Characteizing the shape of histograms statistically...
Jim et al .....
Isn't skewness a method of symmetry/non-symmetry?. However, don't forget
kurtosis, which measures degree of flatness.
..... Andre Francis
On 24 November 2010 08:33, John Bibby <[log in to unmask]> wrote:
> Dear Jim
>
> The standard methods are in terms of symmetric/nonsymettric and
> unimodal/multimodal etc.
>
> But I have often wondered whether there is a _metric_ for symmetry e.g..
> something that takes value 1 when it is completely symmetric and another
> value when it has no symmetry at all (whatever that may mean!).
>
> In the 1970s I worked with a psychologist on a measure for symmetry which
> essentially asked how much of the histogram had to be moved in order to
> make
> it symmetric. It was all pretty straightforward (although the
> distributional
> theory wasn't!), and I don't think anything was ever published.
>
> Relating to modality, one could similarly ask how much of the 'peaks' have
> to be moved into the 'troughs' before it is unimodal.
>
> However the main interest has been to identify mixtures - the story of
> Pearson and Weldon's lobsters is a classic - e.g.
>
>
http://www.google.co.uk/search?sourceid=chrome&ie=UTF-8&q=pearson+weldon+car
apace#q=pearson+weldon+carapace&um=1&ie=UTF-8&tbo=u&tbs=bks:1&source=og&sa=N
&hl=en&tab=wp&fp=101170a721ad585d
>
> I hope that some of this helps.
>
> JOHN BIBBY
>
>
>
> On 24 November 2010 01:23, Jim Silverton <[log in to unmask]> wrote:
>
> > Hello everyone,
> > I was just wondering whether there was a
> > standard way of characterizing the shape of histograms via the use of
> some
> > statistics. For example,
> > I know the shape of the histograms makes particular methods work well.
> But
> > the question is how do I characterize the shape of the histogram
> > statistically?
> >
> > --
> > Thanks,
> > Jim.
> >
> > You may leave the list at any time by sending the command
> >
> > SIGNOFF allstat
> >
> > to [log in to unmask], leaving the subject line blank.
> >
>
> You may leave the list at any time by sending the command
>
> SIGNOFF allstat
>
> to [log in to unmask], leaving the subject line blank.
>
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|