TM1 Forum

It is easy to connect PALO to the statistical package R, using rJava connect for R and Groovy scripting.
Now we finally have proper forecasting and as a bonus great statistical support in the PALO environment.
Using Groovy and Grails, web publishing of your wildest dream statistical graphics are now within reach.

As 'endorsement' InformationBuilders just sent out a press release that they have integrated R under the name Rstat in their Webfocus product.

see:
R: http://www.r-project.org/
rJava: http://rosuda.org/rJava/
Forecasting with R: http://www.robjhyndman.com/
UseR 2008 conference: http://www.statistik.uni-dortmund.de/useR-2008/

Enjoy

Hi Henk

Given that much statistical calculation uses max and min at some point, how do we get past the fact that Palo like TM1 does not seem to have a max or min function?

Is this addressed somehow?

Hi John,

The calculation is done by R.
R is loaded with the (time series) data from PALO and will return the requested forecast series to PALO.
PALO and R are connected through a Groovy/JAVA script.
Forecasting with the R forecasting package goes far beyond what I believe one can do with TM1/PALO/etc
I believe that providing the forecasting quality R offers with PALO calculations will take more than reasonable effort, if possible at all.
For instance based on the time series data, a specific forecasting calculation can automatically be used.

The time series calculation obviously is only a part of a real forecasting solution.
I think the combination PALO+R is great to implement real forecasting.

Beyond forecasting, the calculation and visualisation options the PALO-R connection opens up are great.
The R community provides a whole lot of statistical functions that come free and that are tested.

Regards,

Henk Scholten