New: RapidMiner 4.1

[Ingo Mierswa]

Hello,
 
I just want to announce that a new release of  RapidMiner was published: RapidMiner 4.1.
 
This release is a major bugfix and feature release. In particular, two errors which caused unusual high memory consumptions for some data mining processes were fixed and several new features were added, including 
 

• new 64 bit version for Windows x64 systems
• improved and more comprehensive documentation of all operators inside the programm   
• two different icon sets and look and feels   
• new and improved plotting facilities including different shapes for plotter points   
• the most important preprocessing operators can now produce both views and materialized data   
• several new preprocessing models were added   
• improved wizards for data loading and new wizards for parameter optimizations / iterations   
• performance of several learning schemes improved

 
In total, about 80 bugs were fixed compared to version 4.0 and it is therefore highly recommended to change to version 4.1 instead. In addition, the new features together with a lot of new data mining and analysis operators give you another reasons to try the new version. 
 
The new version and all RapidMiner plugins are  available at 
 
http://rapid-i.com
 
Please post comments and requests in this thread.
 
Cheers,
Ingo

[steffen]

Hello

Shortly after the 4.1 release some bugs appeared in the ParameterIterationOperator. Thank you for fixing it that quick !

What I want to ask: Do you consider a bugfix release ?

greetings

steffen

[Ingo Mierswa]

Hi Steffen,

yes, we do. Actually, the Enterprise Edition already includes a patch which is distributed via an automatic update mechanism (if activated). Please understand that we have to live from something and that we have to provide the patches first to our paying customers - at least in a convenient way, since it is of course also available to all others via CVS (ass you of course know).

On the other hand, there are also good news for our community edition users: we have some new developers and are hence able to reduce the release cycle to 6 to 8 weeks - so the next release will already come end of June. Please understand that we are currently not able to make a release more often since each release (including all tests, find bugs runs, documentation generation and website updates and news / PR) is work for three days. But I am sure that we will reduce the times between releases even further in the future.

Cheers,
Ingo

[B.]

Ingo,

<it is of course also available to all others via CVS>

Some of us are not *nix/Java developers and don't use CVS.  Can you periodically release these updates as a Windows installer?

Thanks.

B.

[Ingo Mierswa]

Hi B.,

Can you periodically release these updates as a Windows installer?

Sure, we will do that. As I said, we try to make new releases every 6 to 8 weeks which I consider quite periodically for a free software and which is more often than it used to be in the past. This speed-up of our release cycle is only possible since we were able to hire some new people - and this of course is only possible since companies decided to work with us or use our services and pay for them. Since each release (and of course fixing the bugs, too) is a lot of work I find it only fair that people paying for these services get them first in in the most convenient way possible.

Together with the growth of Rapid-I we will certainly be able to make even more releases than now and so both user groups, the professional users of the Enterprise Edition and the users of the free Community Edition will faster benefit from the then even more improvements of RapidMiner. And for now I hope that you understand our situation and that you will already be happy about our new release cycle of 6 to 8 weeks for the Community Edition of RapidMiner.

Cheers,
Ingo

[steffen]

Hello

6-8 weeks is great. Beside I am more than amazed of the FREE service you provide for community users, from the small reaction time to requests to (and especially) your presence in the forum.  You always try to help, even if the question is related to very complex experiments ("solve my problems") and even confronted with the x-th request, which normally would lead to a simple and sober: RTFM.

greetings

steffen

[Ingo Mierswa]

Hi Steffen,

thanks for your kind words!

Cheers,
Ingo

[haddock]

Well said Steffen! Couldn't agree more.