Thread Subject: hist with predifined edges

Hallo!

Can you please tell how to get a histogram with the following edges:

edges = [0 200 350 500 800 ceil(max(data))];

It seems that the hist or the histc functions do not work.

Sincerely
I. Botonakis

"Ioannis Botonakis" <i.botonakis_withoutthis_@googlemail.com> wrote in
message news:il8bda$ndo$1@fred.mathworks.com...
> Hallo!
>
> Can you please tell how to get a histogram with the following edges:
>
> edges = [0 200 350 500 800 ceil(max(data))];
>
> It seems that the hist or the histc functions do not work.

HISTC should work -- show the group a SMALL sample of your data and the
exact commands you're using. Then describe what the code does differently
from how you expect/want it to behave.

--
Steve Lord
slord@mathworks.com
To contact Technical Support use the Contact Us link on
http://www.mathworks.com

On Mar 9, 11:55 am, "Ioannis Botonakis"
<i.botonakis_withoutth...@googlemail.com> wrote:
> Hallo!
>
> Can you please tell how to get a histogram with the following edges:
>
> edges = [0 200 350 500 800 ceil(max(data))];
>
> It seems that the hist or the histc functions do not work.
>
> Sincerely
> I. Botonakis

-----------------------------------------------------------------------------------------------
Funny, histc() seemed to work for me. You might need to call tech
support or issue the command "which -all histc" to see if you've
redefined histc. Because with my code it works fine:

% Generate some sample data.
data = 2000*rand(1,100000);
edges = [0 200 350 500 800 ceil(max(data))];
% Take the histogram.
counts = histc(data, edges)
bar(counts);
set(gca,'XGrid','off','YGrid','on');
% Enlarge figure to full screen.
set(gcf, 'Position', get(0,'Screensize'));
% Label axes.
ylabel('Counts', 'FontSize', fontSize);
xlabel('Bins', 'FontSize', fontSize);
% Label the bins (tick marks).
for bin = 1: length(edges)-1
x_labels{bin} = sprintf('%d - %d', ...
edges(bin), edges(bin+1));
end
x_labels{bin+1} = ' ';
set(gca,'XTickLabel', x_labels, 'FontSize', fontSize);

On 11-03-09 10:55 AM, Ioannis Botonakis wrote:

> Can you please tell how to get a histogram with the following edges:
>
> edges = [0 200 350 500 800 ceil(max(data))];
>
> It seems that the hist or the histc functions do not work.

In addition to what the others have said:

If you are going to be using ceil(max(data)) then you should use inf
(infinity) instead. Otherwise you end up with a situation where the last bin
counts values that are equal to the maximum if the maximum _happens_ to be
integral, but where the last bin counts nothing at all if the maximum happens
to have a non-zero fractional part.

"Steven_Lord" <slord@mathworks.com> wrote in message <il8esi$a9k$1@fred.mathworks.com>...
>
>
> HISTC should work -- show the group a SMALL sample of your data and the
> exact commands you're using. Then describe what the code does differently
> from how you expect/want it to behave.
>
> --
> Steve Lord
> slord@mathworks.com
> To contact Technical Support use the Contact Us link on
> http://www.mathworks.com

Dear Mr. Lord,

thank for your answer. I expressed my question incorrectly. I want also the width of the bar columns to be the differnce of the edges. For example:
When the edges are [0 200 350 500 800 ceil(max(data))], then I want that the bin widths are: 200-0=200; 350-200=150; 500-350=150; 800-500 = 300 and so on. Is this possible?

Sincerely
I. Botonakis

On Mar 10, 7:09 am, "Ioannis Botonakis"
<i.botonakis_withoutth...@googlemail.com> wrote:
> thank for your answer. I expressed my question incorrectly. I want also the width of the bar columns to be the differnce of the edges. For example:
> When the edges are [0 200 350 500 800 ceil(max(data))], then I want that the bin widths are: 200-0=200; 350-200=150; 500-350=150; 800-500 = 300 and so on. Is this possible?
>
> Sincerely
> I. Botonakis
---------------------------------------------------
You're going to have to get the handles of the individual bars with
findobj() if you want bars of different widths. I believe that bar()
and the various hist functions all have equal width patches for the
bars, even though the bars represent different width ranges of your
data.