Scilab Home Page | Wiki | Bug Tracker | Forge | Mailing List Archives | Scilab Online Help | File Exchange
ATOMS : Stixbox details
Login with GitLab


Statistics toolbox
(13906 downloads for this version - 65256 downloads for all versions)
A more recent valid version exists: 2.8
Anders Holtsberg
Paris Sud University
Maurice Goursat
Jean-Philippe Chancelier
Serge Steer
Allan Cornet
Michael Baudin
Owner Organization
INRIA,Scilab Enterprises, Universite Paris Sud, and others
Clement DAVID
Michael BAUDIN
Administrator ATOMS
Samuel Gougeon
Creation Date
November 28, 2017
Source created on
Scilab 6.0.x
Binaries available on
Scilab 6.0.x:
Windows 64-bit Windows 32-bit Linux 64-bit Linux 32-bit macOS
Install command
--> atomsInstall("stixbox")
Stixbox is a statistics toolbox which provides distribution functions, datasets,
statistical tests and plotting facilities. 

1) Stixbox for Scilab 6 is available only for Windows and Linux users.
Indeed, Stixbox depends on the distfun and the linalg modules, that are not yet
recompiled for Scilab 6 for MacOS.

2) The following functions have changed:
 * regres(y,X)  => regress(y,X)
 * regresprint(fullstats) => regressprint(fullstats)
 * qqnorm(x) => normplot(x)
 * rexpweib() is removed. Now the distfun module does the job.
   Please replace
    x = rexpweib(n, l, a)
   with :
    distfuna = (1/l)**(1/a);
    distfunb = a;
    x = distfun_wblrnd(distfuna, distfunb, n, 1);

 Please update your existing codes.


 * Datasets
   * getdata : Famous datasets 
 * Graphics
   * Tutorial : Demos of the graphics.
   * boxplot : Draw a box plot
   * bubblechart : Plot a bubble chart
   * bubblematrix : Plot a bubble chart
   * filledbounds : Fill the area between two curves
   * histo : Plot a histogram
   * identify : Identify points on a plot with mouse clicks
   * plotmatrix : Plot an X vs Y scatter plot matrix
   * plotsym : Plot with symbols
   * qqnorm : Normal probability paper
   * qqplot : Create a QQ-plot
   * stairs : Stairstep graph
 * Logistic Regression
   * lodds : Log odds function 
   * loddsinv : inverse of log odd 
   * logitfit : Fit a logistic regression model 
 * Miscellaneous
   * betaln : Logarithm of beta function 
   * cov : Covariance matrix 
   * histocvimse : Cross-validation IMSE of an histogram.
   * histocvsearch : Search optimal CV number of bins in a histogram.
   * ksdensity : Kernel smoothing density estimate
   * ksdensity_kernelcdf : Evaluate kernel for CDF
   * ksdensity_kernelpdf : Evaluate kernel for PDF
   * quantile : Empirical quantile (percentile). 
   * regres : Multiple linear regression
   * regresprint : Print linear regression
   * stixbox_getpath : Returns the path to the current module.
   * summary : five-number summary for a dataset
 * Polynomials
   * polyfit : Polynomial curve fitting 
   * polyval : Polynomial evaluation 
 * Random Numbers
   * rexpweib : Random numbers from the exponential or weibull distributions 
 * Regression
   * cmpmod : Compare linear submodel versus larger one
   * lsfit : Fit a multiple regression normal model
   * lsselect : Select a predictor subset for regression
   * regress : Multiple linear regression
   * regressprint : Print linear regression
 * Resampling
   * T and extra arguments: How to use extra-arguments in T.
   * ciboot : Bootstrap confidence intervals
   * covboot : Bootstrap estimate of the variance of a parameter estimate.
   * covjack : Jackknife estimate of the variance of a parameter estimate.
   * rboot : Simulate a bootstrap resample
   * stdboot : A bootstrap estimate of the parameter standard deviation.
   * stdjack : Jackknife estimate of the standard deviation of a parameter
 * Tests, confidence intervals and model estimation
   * ciquant : Nonparametric confidence interval for quantile
   * kstwo : Kolmogorov-Smirnov statistic from two samples
   * test1b : Bootstrap test that means equals zero
   * test1n : Test that means equals zero (Normal)
   * test1r : Test for median equals 0 using rank test
   * test2n : Tests two normal samples with equal variance
   * test2r : Test location equality of two samples using rank test
Files (2)
[558.60 kB]
Source code archive

[986.18 kB]
OS-independent binary for Scilab 6.0.x
CHANGES (2.5) => 2.6
 * Created new classification functions :
 * Removed bugged and undocumented lsselect.
 Technically replaced by the more powerful stepwiselm.
 * Created new functions :
 * Fixed bug #1665 : boxcoxinverse now produce 
histo failed when the sample is constant and zero.

CHANGES 2.4 => (2.5)
 * Support of Scilab 6 added
 * Fixed bug #1637 : boxcoxinverse now produce 
 NAN when the inverse transformation is not possible.
 * Created Box-Cox transformation, 
 parameter estimation and inverse transformation.
 * Fixed bug #1634
 Improved performance of plotmatrix.
 * Renamed regres into regress into to math Matlab's regress. 
   B = regres(y,X) =>  B = regress(y,X)
 * Renamed regresprint into regressprint in order to be 
   regresprint(fullstats) => regressprint(fullstats)
 * Fixed bug #1633 : regres failed with one single input x
 * Renamed qqnorm into normplot in order to match Matlab's 
  qqnorm(x) => normplot(x)
 * Fixed bug #1635 : qqnorm had wrong labels
 * Added Longley to the datasets.
 * Create getdatanumber() to get the number of available 
 * Fixed bug #1082 : cmpmod had no example
 * Fixed Bug #1631 : removed unnecessary mtlb_length
 * Fixed bug #1550 : plotsym did not manage a single symbol
 * Fixed bug #1627 : layout of plots was not consistent
 * Removed rexpweib.
This fixes bug #1079 : rexpweib had no example.
This fixes bug #1080 : rexpweib should be in distfun.
To update your code, please replace 
// with :
 * Fixed "Warning 1./" in histo.
   regres failed when number of observations was 
   less than the number of parameters
   The help page of regres was not up-to-date.
 * Added an example into the help page of lsselect.
   Fixed example in regresprint.
   Removed cov function, now than a correct implementation 
   is in Scilab.
   Fixed lack of accuracy in polyfit/polyval when 
   problem is ill-conditionned. 
   Added mu argument to standardize (i.e. scale) the data.

News (0)
Comments (8)     Leave a comment 
Comment from Samuel Gougeon -- January 14, 2018, 02:58:05 AM    

This release for Scilab 6 can't be loaded properly, due to a missing Apifun dependency.
Apifun was declared as an explicit dependency up to the 2.4 version, but this is no longer
the case for this 2.6 version, and apparently nested dependencies are not (longer?) well
managed by the atoms installer. So, i get the following error while loading stixbox:

Start Distfun

Start Helptbx
WARNING: Toolbox "apifun" is required and is not installed.

Start Specfun

Start Makematrix

atomsLoad: The file '\loader.sce' from (apifun - 0.4.2) does not exist or is not read
atomsLoad: An error occurred while loading 'apifun-0.4.2':
	exec: Cannot open file \loader.sce.
at line   335 of function        atomsLoad (
C:\Logiciels\scilab\scilab-branch-6.0\modules\atoms\macros\atomsLoad.sci line 351 )
at line     1 of executed string 
It may also be an issue due to the ORDER of declared dependencies. To be investigated.
For the time being, it would be safer to declare again Apifun as an explicit dependency.

Answer from Samuel Gougeon -- June 8, 2018, 09:02:14 PM    
I no longer get any error while loading Stixbox with Scilab 6.0.1 / win7
Comment from Nils Petersen -- April 16, 2018, 09:50:34 PM    
it is not possible to install stixbox on MacOS under scilab version 6.0.1.
Do you have a solution for solving this problem?
Comment from Samuel Gougeon -- June 8, 2018, 09:04:23 PM    
Hello Nils,
May you provide more information about the failure on MacOS, and how you proceeded?
Did you get any error message, etc?
Comment from Julien Chevallier -- September 4, 2018, 01:56:27 PM    

I have the same problem as Nils. 
Here is what the console gives :
--> atomsInstall("stixbox")
at line    62 of function atomsGetVersions ( /Applications/scilab-

line 78 )
at line    96 of function atomsDepTreeFlat ( /Applications/scilab-

line 141 )
at line    86 of function atomsInstallList ( /Applications/scilab-
125 )
at line   233 of function atomsInstall     ( /Applications/scilab- line 249 )

atomsGetVersions: Wrong value for input argument #1: "distfun" must be an
available atoms package 

I looked for the dependencies of stixbox and I think that the problem comes from the fact
that both 
Distfun and Linear Algebra are not available on version 6.0.1 under MacOS.
Answer from Samuel Gougeon -- September 4, 2018, 09:21:46 PM    
Hello Julien,
> .../...
> I looked for the dependencies of stixbox and I think that the problem comes from the
> fact that both Distfun and Linear Algebra are not available on version 6.0.1
> under MacOS.

You are definitely right. 
I am adding a warning about this restriction in the Description's header of the module.
Thanks for your investigation and for having pointed out this issue.
Comment from Anton Gurkov -- October 16, 2018, 05:54:43 PM    
I suggest updating names of the functions regres() and regresprint() to the current 
versions regress() and regressprint() in the description. Probably, it also would be great

to mention fact of this renaming somewhere on this page. For me it was quite tricky to 
figure out why my old code with this function doesn't work anymore :) Thanks!
Answer from Samuel Gougeon -- November 2, 2018, 09:37:44 AM    
Thanks Anton for having pointed out this missing information (buried in the changelog, 
only available after installation).
The changelog for the 2.5+2.6 releases is now published on this page near the related 
file, and as warning in the description.
Please note that qqnorm() has also been renamed, and rexpweib() has been removed (with a 
proper replacement).
Comment from Stéphane MOTTELET -- March 22, 2019, 11:03:03 PM    
Stixbox is claimed to be available for OSX, but the disfun atoms module, which is a
dependency, is not 
available for this platform :

--> atomsInstall stixbox
à la ligne    62 de la fonction atomsGetVersions ( /Applications/scilab-

ligne 78 )
à la ligne    96 de la fonction atomsDepTreeFlat ( /Applications/scilab-

ligne 141 )
à la ligne    86 de la fonction atomsInstallList ( /Applications/scilab-
125 )
à la ligne   233 de la fonction atomsInstall     ( /Applications/scilab- ligne 249 )

atomsGetVersions: Wrong value for input argument #1: "distfun" must be an
available atoms package 

Answer from Samuel Gougeon -- March 23, 2019, 05:26:07 AM    
This is already reported as comment hereabove:

Stixbox description includes a warning about this point:
1) Stixbox for Scilab 6 is available only for Windows and Linux users.
Indeed, Stixbox depends on the distfun and the linalg modules, that are not yet
recompiled for Scilab 6 for MacOS.

Since the module is made only of macros, it is rather artificial to split its binaries 
into four identical copies just renamed and declared according to supported OS.
So, the error is rather on the ATOMS side,
* either to be able to declare OSes even when only macros are involved.
  AFAIR, this wish is already reported, for the TCL cases (as for the /serial/ module)
  that is not supported by MacOSX.
* or/and to able to take dependencies availabilities (maybe also reported.
  to be confirmed)
Comment from Hani Ibrahim -- November 11, 2019, 08:05:14 PM    
Stixbox 2.6 for macOS and Scilab 6
Stixbox does not install in Scilab 6 cause the developer of Distfun and Linalg does not
provide binaries for macOS and Scilab 6.

I compiled both toolboxes and want to share it with the Scilab community. You can download
it at:

Install both toolboxes via -->


Then you can install Stixbox 2.6 with Atom's GUI or commandline as usual.

Have fun
Leave a comment
You must register and log in before leaving a comment.
Login with GitLab
Email notifications
Send me email when this toolbox has changes, new files or a new release.
You must register and log in before setting up notifications.