This is the mail archive of the
gsl-discuss@sourceware.cygnus.com
mailing list for the GSL project.
Re: Does GSL have ANOVA and multiple regression?
- To: Mark Galassi <rosalia@lanl.gov>, GSL discussion list <gsl-discuss@sourceware.cygnus.com>
- Subject: Re: Does GSL have ANOVA and multiple regression?
- From: "G. Anthony Reina" <reina@nsi.edu>
- Date: Mon, 02 Aug 1999 10:06:44 -0700
- Organization: The Neurosciences Institute
- References: <37A5A585.430F2585@nsi.edu> <76u2qinncf.fsf@odie.lanl.gov>
Mark Galassi wrote:
> Anthony> I couldn't find these functions listed so I assume the
> Anthony> answer is "not yet":
>
> You are quite possibly right.
>
> Anthony> Does the GSL library have ANOVA and multiple regression
> Anthony> libraries for C?
>
> Could you mail some pointers to information about these algorithms and
> what problems they solve? Then you might work with people on this
> list to implement them!
ANOVA is the analysis of variance. It compares the variance among groups
of data to see if the data comes from the same population or from
different populations. As an example, say you had three car factories and
wanted to know whether one factory produced more cars than the other two.
An ANOVA could do this.
Multiple regression is finding a line to fit a dataset. So maybe you have
3 independent variables (e.g. age, weight, height) and want to find a
relationship (i.e. a line based on these three variables that fits) to
one depedent variable (e.g. salary).
e.g. salary = b0 + b1 * age + b2 * weight + b3 * height
Good algorithms exist in the IMSL libraries (Visual Numerics,
http://www.vni.com) and MATLAB toolboxes (Mathworks,
http:/www.mathworks.com/access/helpdesk/help/pdf_doc/stats/stats_tb.pdf).
-Tony