Niklas Nebel wrote:
Leonard Mada wrote:
I wish to discuss the following points:
1.) NEED HELP to compile
Have you seen http://wiki.services.openoffice.org/wiki/Windows? If you
have questions about that, try the tools-dev list.
Well, I have never build anything (except for some simple console
applications which I wrote and "run" - NOT really build - with an old VC
6.0). So, I have the option to learn for the next half or so years how
to build OOo (me NOT very enthusiastic), or someone else builds this in
a matter of 2-3 hours and I can use that additional time left (half a
year) to make extensive tests, implement new features and the like ... I
prefer this latter approach as it is this I can do best: test the
statistical functions, search for errors, devise new algorithms.
2.) ANOVA
I'd suggest to concentrate on the existing functions for the moment,
rather than adding new ones.
Well, I noticed that the ANOVA algorithm fits very well in this section.
Like its name says, Analysis of Variance, it is indeed similar to the
Correlation Coefficient and some of the other functions defined in this
particular module. And a couple of users have asked for the
implementation of ANOVA, so maybe it is not so wrong to do it now.
I had a more extensive thought at the one matrix situation, where my
current algorithm split the original matrix in a number of submatrices.
I can think of a better solution now which does NOT require this
sub-splitting (which I did NOT know how to implement, either). It would
speed things a little bit, too, but necessitates splitting the ANOVA
code and adapting the algorithm for this situation. Quite easy to do:
the external LOOP will handle the different variables, while the
internal LOOP will do the rest.
For the "multimatrix input" case, the code would remain unchanged.
Kind regards,
Leonard Mada
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
