The Altair Community is migrating to a new platform to provide a better experience for you. In preparation for the migration, the Altair Community is on read-only mode from October 28 - November 6, 2024. Technical support via cases will continue to work as is. For any urgent requests from Students/Faculty members, please submit the form linked here

Data-Stream-Plugin (formerly Concept-Drift-Plugin)

hagen85hagen85 Member Posts: 18 Contributor II
edited November 2018 in Help
Hi there

I ve downloaded the rapidminer-datastream-4.0beta  plugin and put it into the plugins folder. But it seems there are no new operators available in RapidMiner. Are the corresponding operators already part of RapidMiner 5? If so, what are the names of the available operators for handling concept drift? Are there any sample processes for it?

Thank you and kind regards
Hagen

Answers

  • haddockhaddock Member Posts: 849 Maven
    Hi there,

    Have you seen the series plugin for Rapidminer 5? Rapidminer 4.0 is only dimly remembered, and even less supported!

  • hagen85hagen85 Member Posts: 18 Contributor II
    Thank you for your reply. I have installed the Series Plugin but could not find anything for drift detection. Is there any chance to use adaptive window-sizes in RM?
    Best Regards
    Hagen

    PS: What I have found - and what is pretty cool - are operators for Bagging!! I really start to like that tool.
  • haddockhaddock Member Posts: 849 Maven
    Hi there,

    I seem to remember a sliding window XValidation operator in the Series plugin, that might help...

    Hope so!
  • hagen85hagen85 Member Posts: 18 Contributor II
    Hi haddock

    Thanks for your help. I have already discovered that one. But this operator uses a fixed window, which makes no sense in presence of a stable concepts. Therefore it would be better to have an adaptive window. I just read about a method which is called drift detection method and is used to determine the point in a time-series at which a concept drift has occurred. Basically it monitors the classification error which will decrease over time in the absence of concept drift. The detection has two steps. First a warning(1) and then a point where it is almost certain(2) that a concept drift has occurred. If this happens the data instances between (1) and (2) are used for building a new prediction model. Sounds pretty easy, that is why I thought there might be an implementation for RapidMiner?

    Best regards
    Hagen



  • MALINIMALINI Member Posts: 1 Learner III
    hi where we can find series plugin pl help I am using rapidminer 5
  • hagen85hagen85 Member Posts: 18 Contributor II
    Hi,

    just open  rm and go to help-> update and select the series extension.

    Regards
    Hagen
  • GoceGoce Member Posts: 2 Contributor I
    Hi all,

    I have one question. Can you use Rapid Miner for data steam minig (Algorithms such as VFDT Very Fast Decision Tree)? Is there a tutorial available, anywhere?

    Goce
Sign In or Register to comment.