Due to recent updates, all users are required to create an Altair One account to login to the RapidMiner community. Click the Register button to create your account using the same email that you have previously used to login to the RapidMiner community. This will ensure that any previously created content will be synced to your Altair One account. Once you login, you will be asked to provide a username that identifies you to other Community users. Email us at Community with questions.

[SOLVED] RM 6.1: "Optimize Parameters (Evolutionary)" -> "no data was delivered"

User6813User6813 Member Posts: 2 Contributor I
in Help
System Information
Application Version: RapidMiner Studio 6.1.000 Professional Plus Edition
Machine: MacBook Pro, Retina, 15-inch, Early 2013
Operating System: OS X 10.9.5 (13F34)

Problem

I made two identical processes, with the only difference that one uses the "Optimize Parameters (Grid)" operator, while the other uses the "Optimize Parameters (Evolutionary)" operator instead.

The first one runs correctly, but the second one fails in the "Write Model" operator with the error message "no data was delivered at port".

To be sure, I grepped the two .rmp files for <connect .../> and confirmed that the connections are identical between the two processes.

By the way, I've seen the hint to disable "Parallelize main process" in another forum post that was apparently about the same problem.
But I couldn't find that option anywhere in RM or the documentation. Maybe it's suppressed in my installation.

Here's a zip file containing the processes and data for reproducing the problem:

http://gert-rieger-edv.de/RM_Problem_01/RM_Problem_01_data.zip

And here's a detailed description with screenshots:

http://gert-rieger-edv.de/RM_Problem_01

Answers

  • awchisholmawchisholm RapidMiner Certified Expert, Member Posts: 458 Unicorn
    Hello

    The model output in the Grid search case is almost certainly not the best because the grid search will simply stop at the last combination of parameters and the output will be the last model built, not the best. The same is very likely to be true in the Evolutionary case especially if early stopping is not used.

    The way to solve this is to use the Set Parameters operator combined with the parameter output from the Optimize process. This allows the model to be run again using those parameters.


    regards

    Andrew
  • Marco_BoeckMarco_Boeck Administrator, Moderator, Employee, Member, University Professor Posts: 1,996 RM Engineering
    Hi,

    as you're having a Professional Plus Edition, you can ask our Enterprise Support :)
    Please click here to open a ticket: https://support.rapidminer.com/

    Regards,
    Marco
  • User6813User6813 Member Posts: 2 Contributor I
    Hi Andrew and Marco,
    that solves the mystery, thanks.
Sign In or Register to comment.