Hi Wojciech,
Indeed, I am eagerly awaiting to see the graphics implemented.
A.) In the meantime I have added another example on the
http://wiki.services.openoffice.org/wiki/R_and_Calc_Documentation#Coding_Window
wiki page.
The last example shows a multivariate linear regression model (including
model testing using the ANOVA-test). Should work fine. However, please
be careful when pasting from the wiki, as the wiki-features required to
insert a space between 2 braces, changing '[[' to '['_space_'[' which
breaks the code. One has to convert back these double-braces.
While the model is written for 3 independent variables, one can easily
extend it to more variables (and even add non-linear cross-effects like
x[,2]*x[,3]). One just has to be careful with the indexes of the return
object.
B.) When I have a little bit more time (maybe during the week-end), I
wish to move the R-pages to /Calc/To-Dos/Statistics/... in order to have
the site more structured.
C.) I will think also about various language-design details later on. As
an idea, maybe it would be useful to have a 'print'-method and a
'summary'-method, that outputs exactly the 'print(R-object)', as when
writing
> _my_R_object [ENTER] <- the print method is internally called
... some output
Another method to know the names in the resulting R-object is to call
the function "names('R-object')" which will output all names, e.g.
fit<-lm(x~y+z)
> names(fit)
[1] "coefficients" "residuals" "effects" "rank"
[5] "fitted.values" "assign" "qr" "df.residual"
[9] "xlevels" "call" "terms" "model"
For (S3)-methods, one can use (one has to know the class):
> methods(class="lm")
[1] add1.lm* alias.lm* anova.lm
case.names.lm*
[5] confint.lm* cooks.distance.lm* deviance.lm*
dfbeta.lm*
[9] dfbetas.lm* drop1.lm* dummy.coef.lm*
effects.lm*
[13] extractAIC.lm* family.lm* formula.lm*
hatvalues.lm
[17] influence.lm* kappa.lm labels.lm*
logLik.lm*
[21] model.frame.lm model.matrix.lm plot.lm
predict.lm
[25] print.lm proj.lm* residuals.lm
rstandard.lm
[29] rstudent.lm simulate.lm* summary.lm
variable.names.lm*
[33] vcov.lm*
Non-visible functions are asterisked
One can see, that there is a 'summary.lm'-method, which is internally
called when one writes summary('fit_model').
Hope this helps further. I will also think more thoroughly about further
improvements.
Sincerely,
Leonard
Wojciech Gryc wrote:
Hi Leonard,
Yes, right now the "Launch Rserve" command doesn't work... I have no
clue why, and so far talking with OO developers hasn't yielded
anything useful. I will keep trying!
Worse comes to worst, I'll make a script that launches Rserve along
with a Spreadsheet, though this is a suboptimal solution in my view.
I will add the diagnostics this week. I first want to get image
imports to work, and then I'll definitely do this, along with adding
the tools you mentioned.
THANK YOU SO MUCH for updating the wiki with coding examples. This
really made my day. :) How did you find coding in it? If there's any
way I can make the coding syntax easier to use, or even improve the
user interface, please let me know...
Right now my biggest concern is that you still need to now the R
object structure fairly well to code in the window, but I'm not sure
how to show people what's available as they code. Any ideas?
Thanks,
Wojciech
On 8/14/07, *Leonard Mada* <[EMAIL PROTECTED]
<mailto:[EMAIL PROTECTED]>> wrote:
Hi Wojciech,
nice to see further development.
I had some problems to run the package and did not have time until now
to look into the issue any further. But I installed today v 0.1.4 and,
although I had again problems running it, I was fortunately able,
after
some tweaking, to get it to work. It is really mysteriously, though I
have no clue what was going on (RCONNECT() returned always 0, although
the dialog box kept saying Rserve started. I start it now manually.)
I played a little bit around and by the way, wrote some *coding
window*
examples on
http://wiki.services.openoffice.org/wiki/R_and_Calc_Documentation#Coding_Window
<http://wiki.services.openoffice.org/wiki/R_and_Calc_Documentation#Coding_Window>
.
In addition, I would like to have some menus for the following
R-commands:
[Menu Entry] -> [Sub-Menu]: [R-Syntax]
Count Data
-> Fisher-exact Test: fisher.test({#R1C1:R2C2})
-> Chi-Square Test: chisq.test({#R1C1:R2C2})
where R1C1:R2C2 is the 2x2 contingency table (aka 4 cells are passed;
one can test more than 2 groups, but probably 2x2 contingency
tables are
the most frequent)
Non-Parametric Tests
-> Wilcoxon Rank Sum Test: wilcox.test({#R1C1:RnC1},{#R1:C2:RnC2})
Survival Analysis
-> Kaplan-Meier Curves: a graphic
-> Parametric Model
-> Cox-Proportional Hazards Model
I have updated these on the Stat wiki page (see
http://wiki.services.openoffice.org/wiki/Calc/To-Dos/Statistical_Data_Analysis_Tool#Survival_Analysis
<http://wiki.services.openoffice.org/wiki/Calc/To-Dos/Statistical_Data_Analysis_Tool#Survival_Analysis>)
(with the exception of the Cox-Prop.-Hazards model)
Keep the good work,
Sincerely,
Leonard
P.S. and please implement an:
ABOUT-Function/Menu Entry where one can clearly see the version
of the
package installed
INFO-Function/Menu Entry that displays Rserve() version, path,
and R's
sessionInfo() as a minimum
Wojciech Gryc wrote:
> Hi everyone,
>
> I just wanted to share a very exciting development with you! The
last
> major addition to the package that I promised a few months back was
> graphical output, and today I succeeded in doing this. You can see
> some screenshots on the blog, with a direct link to the post here:
>
> http://www.11-55.org/ooblog/?p=38
>
> I haven't released an updated package for this yet because I'd
like to
> improve upon it. Right now refreshing doesn't work very well,
but I'm
> thinking if I output to an image that then imports into the
> spreadsheet, this might work much better (not sure if I'll get this
> specific feature done soon, though).
>
> Aside from that, the only thing I'm doing is testing the add-on and
> improving the interface. There's a few people who have tried to run
> the add-on and are doing so successfully, so things are looking
good!
>
> Thanks,
> Wojciech
>
--
Five Minutes to Midnight:
Youth on human rights and current affairs
http://www.fiveminutestomidnight.org/
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
