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.

"clustering signals"

lazyfxlazyfx Member Posts: 1 Learner III
edited June 2019 in Help

Hello all,

I am trying to cluster noisy signal data. My data consists of 145 patterns with 150 timepoints each and is very noisy. When a few patterns are averaged together, the waveform appears to be sinusoidal.

My goal is to obtain an unknown number of clusters of these signals that have similar phase shifts. I've tried hierarchical agglomerative clustering with cosine similarity, but I believe that my data is much too noisy. From this, I obtained random clusters with very little difference between them. Any suggestions?

Thanks!

J
Tagged:

Answers

  • haddockhaddock Member Posts: 849 Maven
    Hi there J,

    Firstly, welcome! Secondly, have you tried the Series extension? I suspect it will be just what you need, it is a very nice toolkit for signal scramblers. What's more it is close to the heart of the great Magus Ingo - indeed it is rumoured that he uses it to co-ordinate his crazy dance moves. He actually  has built music classifiers with it ( but probably just to stop tangoing to hip-hop ).

    Again, welcome  ;D

    Extension? Help->Update Rapidminer etc.
  • IngoRMIngoRM Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, Community Manager, RMResearcher, Member, University Professor Posts: 1,751 RM Founder

    but probably just to stop tangoing to hip-hop
    THAT was my intention. But after having realized that this was actually the secret of my moves I stopped this project again  ;D

    Welcome from me as well. Cheers,
    Ingo
Sign In or Register to comment.