R Extension for RapidMiner: Sneak Peek!

[Ingo Mierswa]

Hi,

I just wanted to let you know that the brand-new R extension for RapidMiner will be revealed next week at the RapidMiner conference RCOMM. However, you can have a sneak peak today:

http://rapid-i.com/content/view/219/1/

The web site features a video showing how easy R models and scripts can be integrated into the RapidMiner analysis processes. RapidMiner offers a new R perspective consisting of the known R console together with the great plotting facilities of R. All variables as well as R scripts can be stored in the RapidMiner Repository and used from there which helps to organize the usually large number of scripts. Furthermore, widely used modeling methods are directly integrated as RapidMiner operators as usual.

RapidMiner and R are the most widely used open source solutions for data analysis and it is great that those two programs finally are seamlessly integrated. Let me know what you think!

Cheers,
Ingo

[earmijo]

This is really great news. Those are indeed the two products I use regularly for data analysis. Actually the integration between R and a DM package was the only reason I still had Knime in my desktop.

[Ingo Mierswa]

Glad you like it! I think that indeed many many people already combine RapidMiner and R and exchange data via files. The Extension will now make things much more comfortable and in my opinion even overcomes some issues known from the usual R console for example. I am really looking forward to the release myself!

Cheers,
Ingo

[awchisholm]

A most excellent addition. Now I have an excuse to spend time learning R properly.

Will it be possible to use R to do some pre-processing of example sets that start in Rapid Miner. I'm thinking of feature extraction from value series.

Andrew

[Ingo Mierswa]

Hi,

yes, things like that are easily possible. Have a look at the video at

http://rapid-i.com/content/view/219/1/

Here you can see how a data set is transformed into the R data structure and how data is transformed back into RapidMiner. It's only a matter of a few simple parameters and you can do every data transformation and preprocessing you can imagine (and program with R). I also really like that.

Cheers,
Ingo

[Ghostrider]

Wow, looks very powerful.  When will it be available?

[Ingo Mierswa]

Hi,

the first version will already be released next week!

But we also have already some plans for future versions, many of those we just came up with after a lot of discussion at the RCOMM 2010. One of the major goals for the next version will be to simplify the integration of own pieces of R code as a predefined operator within RapidMiner. Right now, you will have to use the generic "Execute R Script" operator or you will have to edit an XML description if you want something beside the already delivered R operators. That's fine, but in future versions it is likely that you can define the input and output ports including their pre- and postconditions as well as parameters in order to fully integrate complete self-written R operators with a few clicks only.

But for now we are already happy with the upcoming version 

Cheers,
Ingo

[dragoljub]

Eagerly awaiting the update! 

[Ingo Mierswa]

suspense, suspense...