[Scilab-users] robust boxplot within scilab

[A Khorshidi]

David Chèze wrote
> actually it's the two modules I've already evaluated on this topic.  My
> test case is a set of 83 series of 52 samples as 52x83 matrix, not
> normalized, i want to have an outlook of this data.
> Both NAN and CASCI failed to produce comprehensive plot.

I've tried a similar testcase and I see no problem.

exec("nan_boxplot.sci",-1)
data = 100*rand(85,52);
scf(10001); nan_boxplot(data')
//dat = data($:-1:1);
//scf(10002); nan_boxplot(dat)

Graphic_window_number_10001.png
<http://mailinglists.scilab.org/file/n4031313/Graphic_window_number_10001.png>  


David Chèze wrote
> I'm running stixbox (statpack) as a basis for statistic module therefore
> NAN or CASCI might be confused if some functions have the same name.

You can use stixbox and NAN together because the name of all the NAN's
functions started with "nan_". 


David Chèze wrote
> Looking into boxplot.sci from CASCI (I've not installed the whole
> module),this is the case of the "quantile" function.
> It was not my aim but I'll try to adapt it just for boxplot feature.

BTW, I suggest you modify NAN's function for your target instead of CASCI
one, because nan_boxplot() has better visualization as well as it considers
each column as a separate data set.


Regards, 
Mehran
_






--
View this message in context: http://mailinglists.scilab.org/robust-boxplot-within-scilab-tp4031307p4031313.html
Sent from the Scilab users - Mailing Lists Archives mailing list archive at Nabble.com.